An Omega-3-rich Anti-inflammatory Diet Improved Widespread Allodynia and Worsened Metabolic Outcomes in Adult Mice Exposed to Neonatal Maternal Separation

Abstract

Inflammation plays a key role in the progression and maintenance of chronic pain, which impacts the lives of millions of Americans. Despite growing evidence that chronic pain can be improved by treating underlying inflammation, successful treatments are lacking and pharmaceutical interventions are limited due to drug side effects. Here we are testing whether a ‘healthy human’ diet (HHD), with or without anti-inflammatory components (HHAID), improves pain-like behaviors in a preclinical model of chronic widespread hypersensitivity induced by neonatal maternal separation (NMS). The HHD and HHAID are isocaloric and macronutrient-matched, have a low glycemic index, and fat content (35 kcal%) that is high in omega-3 fatty acids, while only the HHAID includes a combination of key anti-inflammatory compounds, at clinically relevant doses. Mice on these diets were compared to mice on a control diet with a macronutrient composition commonly used in rodents (20% protein, 70% carbohydrate, 10% fat). Our results demonstrate a benefit of the HHAID on pain-like behaviors in both male and female mice, despite increased caloric intake, adiposity, and weight gain. In female mice, HHAID specifically increased measures of metabolic syndrome and inflammation compared to the HHD and control diet groups. Male mice were susceptible to worsening metabolic measures on both the HHAID and HHD. This work highlights important sexual dimorphic outcomes related to early life stress exposure and dietary interventions, as well as a potential disconnect between improvements in pain-like behaviors and metabolic measures.

INTRODUCTION

Chronic pain conditions affect 40–116 million Americans (Dahlhamer et al., 2018; Pitcher et al., 2019), costing an estimated $635 billion annually in health care costs and lost work productivity (Gaskin and Richard, 2012). Sustained, low-grade inflammation is thought to be a contributing factor in many chronic pain disorders (Zhang and An, 2007). Unresolved inflammation following an acute injury can contribute to the development of chronic pain through the continued release of prostaglandins, bradykinin, and proinflammatory cytokines and chemokines that activate nociceptors and drive peripheral sensitization (Abbadie et al., 2009; Gold and Gebhart, 2010; Ren and Dubner, 2010; Schaible et al., 2010). Increased visceral fat accumulation, a key hallmark of obesity and metabolic disorder, also leads to chronic low-grade inflammation (Gregor and Hotamisligil, 2011) and has been postulated to underlie neuroimmune mechanisms that contribute to chronic pain (Eichwald and Talbot, 2020). This is clinically important as obesity and chronic pain are known comorbid disease conditions, despite a paucity of work examining if these conditions are mechanistically linked.

Pharmacological-based treatment of the inflammatory component of chronic pain in either lean or obese states has been limited due to harmful side effects (Payne, 2000; Kidd and Urban, 2001), which has led to research on the effectiveness of diet-based interventions to lower inflammation (Tick, 2015). A diet with increased omega-3 fatty acids, such as flaxseed and soybean oils, has been linked to long-term health benefits such as improved weight management and cardiovascular, immune, and neuronal function (Swanson et al., 2012). Increased omega-3 fatty acid consumption can also reduce headache pain in chronic headache sufferers (Ramsden et al., 2013) and joint pain in individuals with rheumatoid arthritis (Goldberg and Katz, 2007). A low glycemic index diet benefits weight control by promoting satiety as well as fat oxidation rather than carbohydrate oxidation (Brand-Miller et al., 2002). Low glycemic index foods are digested and absorbed at a slower rate, which prevents a quick spike in blood glucose and maintains insulin sensitivity. Migraineurs consuming a low glycemic diet saw a reduction in frequency of migraine attacks (Evcili et al., 2018). Compounds including epigallocatechin gallate (EGCG), sulforaphane, resveratrol, curcumin, and ginseng can be added to a diet for additional anti-inflammatory benefit. EGCG is the most abundant polyphenol found in green tea (Higdon and Frei, 2003) and has been shown to increase mechanical and thermal pain thresholds in a model of chronic constriction injury (CCI) (Kuang et al., 2012). Sulforaphane is found in cruciferous vegetables (Guerrero-Beltrán et al., 2012; Lee et al., 2014) and, like EGCG, prevented the development of CCI-induced mechanical and thermal hypersensitivity (Wang and Wang, 2017). Resveratrol is a natural polyphenol and phytoalexin with anti-inflammatory and antioxidant effects (Rocha-González et al., 2008) and was shown to attenuate thermal hyperalgesia in a mouse model of diabetic neuropathy (Sharma et al., 2007). Curcumin is a bioactive polyphenol found in turmeric (Strimpakos and Sharma, 2008). Clinically, curcumin has been effective in the treatment of knee osteoarthritis (Kuptniratsaikul et al., 2014) and preclinically, prevented the development of thermal and mechanical paw hypersensitivity in CCI (Di et al., 2014). Ginseng is a root that contains pharmacological compounds called ginsenosides (Zhang et al., 2017) and prevented pain-like behaviors following capsaicin injection into the hind paw (Nah et al., 2000). There is growing evidence that these individual anti-inflammatory compounds can improve aspects of chronic pain, however, these studies are limited (Nah et al., 2000; Sharma et al., 2007; Kuang et al., 2012; Di et al., 2014; Kuptniratsaikul et al., 2014; Wang and Wang, 2017; Silva et al., 2019) and additional studies using a combination of these compounds at physiologically relevant doses are warranted (Totsch et al., 2018).

In the present study, we determined the effect of what would be considered a healthy human diet (HHD) containing fat high in omega-3 fatty acids and a low glycemic index on early life stress-associated pain-like behavior and measures of metabolic syndrome in our mouse model of neonatal maternal separation (NMS). The NMS model exhibits evidence of persistent low-level inflammation displayed as urogenital hypersensitivity associated with local neuroimmune activation (Pierce et al., 2014, 2016, 2018; Fuentes et al., 2017), altered glucocorticoid production and receptor expression within the limbic system and hypothalamic-pituitary-adrenal (HPA) axis (Pierce et al., 2014, 2016, 2018; Fuentes et al., 2017), and increased susceptibility to obesity on both a chow and high fat/high sucrose diet (Eller et al., 2020). Due to the role of inflammation in chronic pain and obesity, we also investigated the effect of the HHD with additional anti-inflammatory components (HHAID) on outcomes related to widespread hypersensitivity and metabolic syndrome. The HHAID used in the present study, developed from Totsch et al., (Totsch et al., 2018), contained EGCG, sulforaphane, resveratrol, curcumin, ginseng and the same fat sources and glycemic index as the HHD. While we found that the HHAID effectively reduced NMS-induced urogenital and widespread hypersensitivity, the diet negatively impacted weight gain and adiposity, as well as metabolic health. The HHD had no effect on NMS-induced hypersensitivity and worsened metabolic measures in male mice.

EXPERIMENTAL PROCEDURES

Animals

All experiments were performed on male and female C57Bl/6 mice (Charles River, Wilmington, MA) born and housed in the Research Support Facility at the University of Kansas Medical Center. Mice were housed at 22 °C on a 12-h light cycle (600 −1800 h) and received water and food ad libitum. All research was approved by the University of Kansas Medical Center Institutional Animal Care and Use Committee in compliance with the National Institute of Health Guide for the Care and Use of Laboratory Animals.

Neonatal maternal separation (NMS)

Pregnant C57Bl/6 dams were delivered to the animal facility between 14 and 16 days gestation. Litters were divided equally into NMS and naïve groups. From postnatal day 1 (P1) until P21, NMS pups were removed en masse and placed in a clean glass beaker with bedding from their home cage for 180 min (11 am–2 pm). The beaker was placed in an incubator maintained at 33 °C and 50% humidity. Naïve mice remained undisturbed in their home cage except for normal animal husbandry. All mice were weaned on P22 and pair-housed with same sex litter mates and ad libitum access to water and a control diet composed of 20% kcal protein, 70% kcal carbohydrate (3.5% sucrose), and 10% kcal fat (Research Diets, Inc. New Brunswick, NJ, D17072402; Table 1).

Healthy human diet (HHD) and healthy human anti-inflammatory diet (HHAID)

Studies comparing the HHD (D17072403, Research Diets, Inc., New Brunswick, NJ) and HHAID (D17072401, Research Diets, Inc.) to control diet were carried out in two separate cohorts of mice. In both studies, half of the female and male naïve and NMS mice were placed on either the HHD or the HHAID, while the remainder continued on the control diet. The compositions of the individual diets are found in Table 1.

Perigenital mechanical sensitivity

Perigenital mechanical withdrawal threshold was assessed every 4 weeks starting after 4 weeks on the HHD or HHAID. For 2 days prior to the test day, mice were acclimated to a soundproof room for 30 minutes and then placed into individual clear plastic chambers (11 × 5 × 3.5 cm) on a wire mesh screen elevated 55 cm above a table for 30 min. Additionally, the perivaginal area of female mice was shaved on the first day of acclimation. On the test day, mice were acclimated to the soundproof room for 30 min and then placed on the table for 30 min. The up-down method was performed to test mechanical sensitivity using a standard set of Semmes-Weinstein monofilaments (1.65, 2.36, 2.83, 3.22, 3.61, 4.31, 4.74 g; Stoelting, Wood Dale, IL) (Dixon, 1980; Chaplan et al., 1994). Beginning with the 3.22 g monofilament, mice received a single application to either the scrotum or perivaginal area. A negative response was followed by the next larger filament and a positive response (considered a brisk jerk or licking the probed area) was followed by the next smaller filament. The experimenter continued to move up or down the series, depending on the previously elicited response, for an additional four applications after the first positive response was observed fora minimum of five ora maximum of nine total monofilament applications. The value in log units of the final monofilament applied in the trial series was used to calculate 50% g threshold for each mouse (Chaplan et al., 1994).

Hindpaw mechanical sensitivity

Hindpaw mechanical sensitivity was assessed every 4 weeks starting after 5weeks on the HHD or HHAID. On the test day, mice were acclimated to a soundproof room for 30 min and then placed into individual clear plastic chambers (11 × 5 × 3.5 cm) on a wire mesh screen elevated 55 cm above a table for 30 min. An electronic von Frey device (IITC Life Science Inc. Woodland Hills, CA) was used to measure hindpaw withdrawal threshold. A semi-flexible tip filament was applied to the hindpaw and the force that elicited a withdrawal was recorded from the electronic device. The filament was applied six times to each mouse and the highest and the lowest value for each mouse were excluded. Therefore, an average of 4 measurements/mouse was quantified.

Nest building test

At 1 h before the start of the dark phase (5 pm), mice were individually placed into clean cages containing no environmental enrichment outside of a 3 g nestlet square. Seventeen hours later (10 am), the mice were returned to their home cages and the nests were photographed and intact nestlet pieces were weighed. Two blinded experimenters scored the nests based on a 1–5 scale according to previous publications (Deacon et al., 2002; Deacon, 2012) and the average of their scores is reported here.

Body weight, intake, and feed efficiency

Energy intake (per pair) and body weight were measured weekly. Feed efficiency was calculated as weight gained/calories consumed per pair of mice per week. An average feed efficiency was quantified from weekly feed efficiency.

Body composition analysis

Every 4 weeks, mice were weighed and placed in an EchoMRI 2015 (EchoMRI LLC, Houston, TX) to quantify lean mass and fat mass. At time of sacrifice, mice were overdosed with inhaled isoflurane, the epididymal/periovarian and retroperitoneal fat pads were excised and weighed. These two fat pad weights were summed to calculate total visceral fat.

Fasting insulin

After 18–19 weeks on the HHD or HHAID, fasting insulin level was measured. Following a 6-hour fast, blood was collected via tail-clip, placed on ice for 1 h, and centrifuged at 10,000 rpm for 10 min. Serum was collected and frozen until analysis using an insulin ELISA kit (80-INSMS-E01, ALPCO, Salem, NH) according to the manufacturer’s instructions.

Glucose tolerance test

After 18–19 weeks on the HHD or HHAID, a glucose tolerance test was carried out. Following a 6-hour fast, mice were given an IP injection of glucose at 1 g/kg body weight. Blood was collected via tail clip immediately prior to the glucose injection and 15, 30, 60, and 120 min thereafter and blood glucose concentrations were measured by colorimetric assay (PGO enzyme preparation and dianisidine dihydrochloride, Sigma-Aldrich, St. Louis, MO).

HOMA-Ir

To calculate HOMA-IR we used the formula: fasting insulin (mU/L) * fasting glucose (mg/dL)/405.

Corticosterone

After 18–19 weeks on the HHD or HHAID, mice were sacrificed during the early half of the light cycle (8:00 am–11:00 am) and trunk blood was collected. Serum was removed and frozen until analysis using a corticosterone ELISA kit (55-CORMS-E01, ALPCO, Salem, NH) according to the manufacturer’s instructions

mRNA extraction and RT-PCR

Epidydimal and periovarian adipose tissue was dissected, weighed, and immediately frozen in liquid nitrogen, and stored at −80 °C. Frozen tissue was then crushed (Cellcrusher, Portland, OR) and total RNA was isolated using QIAzol Lysis Reagent and the RNeasy Lipid Tissue Mini Kit (Qiagen, Valencia, CA). The concentration and purity were determined using NanoDrop 2000 (Thermo Fisher Scientific, Wilmington, DE) and cDNA was synthesized from total RNA using the iScript cDNA Synthesis Kit (Bio-Rad, Hercules, CA). Quantitative RT-PCR was performed using SsoAdvanced SYBR Green Supermix (Bio-Rad) and a Bio-Rad iCycler IQ real time PCR system with indicated 20 μM primers (Table 2; Integrated DNA Technologies, Coralville, IA). To reduce variability due to fluctuations in baseline fluorescence, the raw PCR data was imported to the LinRegPCR software and PCR efficiency values were derived for each individual sample. Threshold cycle values were subtracted from that of the housekeeping gene PPiB and the fold change over Naïve-Control was calculated using the Pfaffl method (Pfaffl, 2001)

Statistical analyses

Comparisons were made between the control groups from the HHD and HHAID cohorts based on sex and NMS status and no significant differences were observed for body weight, body fat, food intake, feed efficiency, fasting insulin, serum glucose levels, serum corticosterone, or RT-PCR; therefore, the control groups were combined for statistical comparisons. Area under the curve (AUC) measurements were calculated using the trapezoidal rule. Calculations were made in Excel (Microsoft, Redmond, WA) and normality of distribution was tested using Shapiro-Wilk’s test (p < 0.05). Non-normally distributed data were log transformed prior to statistical analysis. Statistical analyses were performed using GraphPad Prism 8 (GraphPad, La Jolla, CA) or IBM SPSS Statistics 25 (IBM Corporation, Armonk, NY). Differences between groups were determined by two- or three-way ANOVA, with or without repeated measures (RM) and Fisher’s LSD posttest, as denoted in the figure legends. Statistical significance was set at p < 0.05. Statistical analyses of mechanical sensitivity were validated by using a Generalized Estimating Equation (GEE) framework. For hindpaw sensitivity, the following model was created for each diet/sex: $Hindpaw{}_{ij}={\alpha }_0+{\alpha }_{1}week{}_{ij}+{\alpha }_{2}week{}_{ij}^2+{\alpha }_{4}stress{}_{ij}+{\epsilon }_{ij}$ and for perigenital sensitivity the following model was created for each diet/sex: log $\log \left(Perigenital{}_{ij}\right)={\alpha }_0+{\alpha }_{1}wee{k}_{ij}+{\alpha }_{2}wee{k}_{ij}^2+{\alpha }_{4}stress{ }_{ij}+{\epsilon }_{ij}$. To fit these models, a Markov structure was employed for the working correlation matrix. For posttest analyses, each subgroup was compared in terms of area under the curve. Data are displayed as box-and-whisker plots representing the median and the 25th and 75th quartiles, as well as individual data points, when feasible, otherwise mean ± standard error of the mean are shown for repeated measures.

RESULTS

HHAID, but not HHD, reduces perigenital mechanical sensitivity

A significant effect of diet was observed on perigenital withdrawal threshold in both female and male mice (Fig. 1A, C). In naïve and NMS female mice, the HHD had no significant effect. However, the HHAID resulted in a significantly higher cumulative perigenital withdrawal threshold compared to control-fed and HHD-fed mice, indicating that HHAID reduced perigenital mechanical sensitivity in female mice (Fig. 1B). In male mice, the HHD significantly lowered the cumulative perigenital withdrawal threshold in NMS mice compared to control fed mice. HHAID-fed naïve and NMS male mice both exhibited significantly higher mechanical withdrawal thresholds compared to control-fed and HHD-fed mice. Taken together, the HHD selectively increased perigenital mechanical sensitivity in NMS male mice while the HHAID reduced perigenital mechanical sensitivity in all mice, regardless of sex or NMS exposure.

Fig. 1.

Impact of NMS and diet on perigenital mechanical withdrawal threshold. Perigenital withdrawal threshold was measured every 4 weeks in female (A) and male (C) mice and the area under the curve (AUC) was calculated to produce a cumulative threshold (B, D). (A) In female mice, there was a significant effect of diet over the 16 weeks (p < 0.0001). (B) AUC measurements also revealed a significant effect of diet (p = 0.0005) with HHAID-fed mice having significantly higher thresholds than HHD- or control-fed mice, regardless of NMS status. (C) In males, there was a significant effect of diet (p < 0.0001) over the 16 weeks. (D) AUC measurements also revealed a significant effect of diet (p < 0.0001) with HHAID-fed mice having significantly lower withdrawal thresholds compared to HHD- and control-fed for both naïve and NMS mice. Additionally, NMS-HHD mice had significantly greater withdrawal thresholds than both NMS-Control and Naïve-HHD mice. δ denotes a significant effect of diet, three-way RM ANOVA (A, C) or two-way ANOVA (B, D). ††p < 0.01 HHD vs. control, &, &&, &&&&p < 0.05, 0.01, 0.0001 HHAID vs. control, #, ####p < 0.05, 0.0001 HHAID vs. HHD, Fisher’s LSD. n = 6–16 per group.

HHD has no effect, while HHAID attenuates hindpaw allodynia only in female NMS mice

In female mice, there was a significant impact of NMS and time on hindpaw mechanical withdrawal threshold both across all experimental time points (Fig. 2A) and of NMS on the cumulative measurements (Fig. 2B). Female NMS-control and -HHD mice had significantly lower cumulative hindpaw withdrawal thresholds compared to naïve-control and -HHD mice, respectively (Fig. 2B). NMS-HHAID mice had a modest increase in hindpaw withdrawal threshold that was not significantly different from either naïve-HHAID, NMS-control, or NMS-HHD mice (Fig. 2B). In male mice, there was a significant impact of time and a trend toward an NMS effect on hindpaw withdrawal thresholds both across the experimental time points (p = 0.081, Fig. 2C) and for the cumulative measurements (p = 0.079, Fig. 2D). No effect of diet was observed on hindpaw mechanical withdrawal threshold in male mice, regardless of NMS status. Together, these data suggest that NMS had a negative impact on hindpaw sensitivity in female mice, and this NMS-induced decrease was attenuated by the addition of the HHAID diet.

Fig. 2.

Impact of NMS and diet on hindpaw mechanical withdrawal threshold. Hindpaw mechanical withdrawal threshold was measured every 4 weeks (A, C) and the area under the curve (AUC, (B, D)) was calculated to produce a cumulative threshold. In female mice, there was an overall significant effect of NMS (p = 0.001) and time (p < 0.0001) across all experimental time points (A), and of NMS for the cumulative measurement ((B), p = 0.001). (B) Both NMS-Control and NMS-HHD female mice had significantly lower withdrawal thresholds compared to their naïve counterparts. (C, D) In males, there was a significant effect of time (p = 0.0069) and a trend toward an NMS effect (p = 0.081), but no effect of diet. § and τ denote a significant impact of NMS or time, respectively, three-way RM ANOVA (A) or two-way ANOVA (B). *, ***p < 0.05, 0.001 vs. same-diet naïve, Fisher’s LSD posttest. n = 6–16 per group.

HHD has no effect, while HHAID worsens nest building scores in female NMS mice

Nest building was evaluated at the end of the experiment to determine the effect of NMS and diet on a measure of anhedonia. In the females, there was a significant impact of diet on nest building (Fig. 3A) evidenced as lower scores in NMS-HHAID mice, compared to naïve-control mice. There were no observed NMS or diet effects on nest building in male mice (Fig. 3B). Examples of nests with scores ranging from 1.5 to 5 are shown in Fig. 3C. These data suggest that HHAID in female NMS mice increases anhedonic behavior.

Fig. 3.

Impact of NMS and diet on nest building behavior. (A) In female mice, there was a significant overall effect of diet (p = 0.048) on nest quality. NMS-HHAID mice had a significantly lower nest score compared to naïve-control mice. (B) Nest quality was not affected by diet or NMS in male mice. (C) Examples of nests with scores ranging from 5 (left) to 1.5 (right) are shown. δ denotes significant effect of diet, two-way ANOVA. *p < 0.05 vs. naïve-control mice, Fisher’s LSD posttest. n = 6–16 per group.

HHAID increases body weight and fat gain in both female and male NMS mice, whereas HHD only impacts male mice

In females, there was a significant impact of NMS, diet, and time on body weight across the entire experiment (Fig. 4A) and at time of euthanization (Fig. 5A). After 18 weeks on the diets, HHD- and HHAID-fed female mice were significantly heavier than control-fed, regardless of NMS status (Fig. 5A). In addition, NMS-HHAID female mice were significantly heavier than naïve-HHAID mice. Only diet had a significant impact on body fat percentage across all time points (Fig. 4B) and at euthanization (Fig. 5B). Naïve-HHAID and NMS-HHAID mice had significantly greater body fat percentage compared to naïve-control and NMS-control mice, respectively (Fig. 5B). NMS-HHAID mice also had a significantly greater body fat percentage compared to NMS-HHD mice (Fig. 5B). The weight of the periovarian fat pad was significantly impacted by diet and was significantly heavier in NMS-HHAID female mice compared to either NMS-control or NMS-HHD mice (Fig. 5C).

Fig. 4.

Impact of NMS and diet on body weight and composition. (A) In female mice, there was a significant overall effect of NMS (p = 0.007), diet (p = 0.006), and time (p < 0.0001) on body weight across 18 weeks. NMS-HHAID mice weighed significantly more than NMS-control (weeks 7, 9, and 14–18) and NMS-HHD (weeks 14–18) mice. (B) Body composition in females was significantly impacted by diet (p < 0.0001) and time (p < 0.0001). NMS-HHAID mice had significantly higher body fat percentage compared to NMS-control mice (every timepoint) and NMS-HHD mice (all but 12 weeks). At 18 weeks on the diet, naïve-HHAID mice had significantly higher body fat percentage than naïve-control mice. (C) In male mice, there was a significant overall effect of diet (p < 0.0001) and time (p < 0.0001)on body weight. NMS-HHAID (weeks 5–18) and-HHD (weeks 10–18) mice weighed more than NMS-control mice. Similarly, naïve-HHAID (weeks 14–18) and -HHD (weeks 10–18) mice weighed more than naïve-control mice. (D) Body composition analyses in male mice found a significant overall effect of diet (p < 0.0001) and time (p < 0.0001) on body fat percentage. NMS-HHAID mice had significantly higher body fat percentage compared to NMS-control mice (weeks 8–18). At week 16, naïve-HHAID mice had significantly higher body fat percentage compared to naïve-control mice and naïve-HHD had significantly higher body fat percentage compared to naïve-control (weeks 12–18). §, δ, and τ denote significant effects of NMS, diet, and time, respectively, three-way RM ANOVA. *p < 0.05 vs. same-diet naïve, †p < 0.05 HHD vs. control, &p < 0.05 HHAID vs. control, #p < 0.05 HHAID vs. HHD, Fisher’s LSD posttest. n = 6–16 per group.

Fig. 5.

The impact of NMS and diet on end-of-study body weight, body fat percentage, and gonadal fat weight. (A) Final body weight in female mice was significantly impacted by NMS (p = 0.044) and diet (p < 0.0001). HHD- and HHAID-fed mice were significantly heavier than control-fed mice, regardless of NMS status. NMS-HHAID female mice were also significantly heavier than naïve-HHAID mice. (B) Final body fat percentage was significantly impacted by diet (p < 0.0001), specifically in HHAID-fed mice. (C) Periovarian fat weight was significantly impacted by diet (p < 0.0001) with NMS-HHAID mice having significantly heavier fat pads than NMS-control and NMS-HHD mice. In male mice, final body weight (D), body fat percentage (E), and epididymal fat pad weight (F) were all significantly impacted by diet (p < 0.0001) with HHD- and HHAID-fed mice having significantly higher values compared to control-fed mice, regardless of NMS status. § and δ denote significant effects of NMS and diet, respectively, two-way ANOVA. *p < 0.05 vs. same-diet naïve, †, ††, †††, †††† < 0.05, 0.01, 0.001, 0.0001 HHD vs. control, &, &&, &&&, &&&&p < 0.05, 0.01, 0.001, 0.0001 HHAID vs. control, #p < 0.05 AID vs. HHD, Fisher’s LSD posttest. n = 6–16 per group.

In male mice, only diet and time significantly impacted body weight (Fig. 4C) and body fat percentage (Fig. 4D) across all time points. After 18 weeks on the diets, male HHD- and HHAID-fed mice had significantly greater body weight (Fig. 5D) and body fat percentage (Fig. 5E) compared to control-fed mice, regardless of NMS status. Epididymal fat pad weight was significantly impacted by diet and was significantly heavier in HHD- and HHAID-fed male mice compared to control-fed, regardless of NMS status (Fig. 5F). Together these data suggest that the HHAID had a greater effect on driving adiposity in NMS female mice, whereas both HHD and HHAID drove weight and fat gains in naïve and NMS male mice to a similar extent.

HHD and HHAID impact food intake and feed efficiency

In females, calories consumed were significantly impacted by diet across the duration of the experiment (Fig. 1A). Female naïve and NMS mice fed either HHD or HHAID consumed more calories/pair compared to control-fed mice at nearly every time point measured. Feed efficiency in female mice was also significantly impacted by diet, however, despite being equal in caloric density and similarly impacting caloric intake, HHAID-fed female mice had a significantly increased feed efficiency compared to HHD-fed mice, regardless of NMS status (Fig. 1B). In males, caloric intake was significantly impacted by diet (Fig. 1C) with HHD- and HHAID-fed mice consuming significantly more calories than control-fed mice at nearly every time point, regardless of NMS status. Feed efficiency in male mice was also significantly impacted by diet, however, unlike in the female mice, there was no significant difference in feed efficiency between HHD- and HHAID-fed mice (Fig. 1D). These data show that HHD and HHAID increased caloric intake in both male and female mice and that both diets increased feed efficiency in male mice with only HHAID increasing feed efficiency in female mice.

Fig.6.

Effect of diet and early life stress on food intake and feed efficiency. (A) In female mice, there was a significant effect of diet (p = 0.042) on caloric intake. NMS and naïve mice fed either HHAID or HHD consumed more calories than their control-fed counterparts throughout the experiment. (B) Feed efficiency in female mice was significantly impacted by diet (p = 0.0004) with naïve-HHAID and NMS-HHAID mice having significantly higher feed efficiencies compared to their HHD-fed counterparts, despite being calorically identical. (C) In male mice, there was a significant overall effect of diet (p = 0.007) on caloric intake with HHAID- and HHD-fed mice consuming more calories than control-fed mice at every time point of the study. (D) There was also a significant effect of diet (p = 0.046) on feed efficiency in male mice, but no statistical difference between HHD- and HHAID-fed mice. δ denotes a significant effect of diet, three-way RM ANOVA (A, C) or two-way ANOVA (B, C). †p < 0.05 HHD vs. control, &p < 0.05 HHAID vs. control, #, ##p < 0.05, 0.01 HHAID vs. HHD, Fisher’s LSD posttest. n = 3–8 pairs per group.

HHD and HHAID differentially affect serum glucose and insulin levels in male and female mice

Diet significantly impacted serum glucose levels in female mice during a glucose tolerance test (GTT), such that NMS-HHAID female mice had significantly higher serum glucose levels at 15, 30, and 60 min into the GTT (Fig. 7A), and overall (Fig. 7B) compared to NMS-control female mice. NMS-HHD female mice also had significantly higher GTT AUC measurements compared to NMS-control mice (Fig. 7B). Female naïve-HHAID mice also had significantly higher serum glucose levels at 30 min into the GTT (Fig. 7B) and overall (Fig. 7C) compared to naïve-control female mice. No significant impact of NMS or diet was observed on fasting serum insulin levels (Fig. 7C) or HOMA-IR measurements (Fig. 7D) in female mice.

Fig. 7.

Impact of diet and early life stress on glucose tolerance, fasting insulin, and HOMA-IR. (A) In female mice, there was a significant effect of diet (p < 0.001) on glucose tolerance with female NMS-HHAID mice having significantly higher serum glucose compared to NMS-Control mice at 15, 30, and 60 min and higher than NMS-HHD mice at 30 min. Additionally, Naïve-HHAID mice had a higher serum glucose level at 30 min compared to Naïve-Control mice. (B) Area under the curve (AUC) measurements revealed a significant diet effect (p = 0.0003) with both naïve-HHAID and NMS-HHAID mice having significantly higher serum glucose compared to their control counterparts. NMS-HHD mice were also significantly higher than NMS-Control. There was no significant effect of diet or NMS on fasting serum insulin levels (C) or calculated HOMA-RI (D). (E) In male mice, a significant effect of diet (p = 0.004) was observed on glucose tolerance in male mice with naïve-HHD mice having significantly higher serum glucose levels at 30 and 60 minutes compared to naïve-control mice. (F) AUC measurements were also significantly impacted by diet (p = 0.005) with naïve-HHD mice having higher serum glucose levels than naïve-control mice. A significant impact of diet was observed on fasting serum insulin levels (p = 0.040) and HOMA-IR (p = 0.032) in male mice with naïve-HHAID and naïve-HHD mice having significantly higher levels than naïve-control mice. δ denotes a significant effect of diet, three-way RM ANOVA (A, E) or two-way ANOVA (B-D, F-H). †, ††p < 0.05, 0.01 HHD vs. control, &, &&&p < 0.05, 0.001 HHAID vs. control, #p < 0.05 HHAID vs. HHD, Fisher’s LSD posttest. n = 6–16 per group.

Diet significantly impacted serum glucose levels in male mice during a GTT, predominantly in naïve mice (Fig. 7E, F). Naïve-HHD male mice had a significantly higher serum glucose level compared to male naïve-control mice at 30 and 60 min into the GTT (Fig. 7E). There was a significant effect of diet on male GTT AUC with naïve-HHD male mice having significantly higher GTT AUC measurements compared to male naïve-control mice (Fig. 7F). Diet also significantly impacted fasting serum insulin levels (Fig. 7G) and HOMA-IR measurements in male mice (Fig. 7H), such that naïve-HHD and naïve-HHAID mice were significantly higher than naïve-control mice. No significant diet effects were observed in NMS male mice. Overall, these data suggest that NMS renders female mice more susceptible to HHD- and HHAID-induced impairments in glucose tolerance compared to female naïve mice, whereas male mice, regardless of NMS exposure, have worsened outcomes due to HHD and HHAID, with these effects likely being driven by the increased fat content in both diets.

Diet and NMS differentially affect serum corticosterone levels in female and male mice

Serum corticosterone levels were significantly impacted by diet in both female (Fig. 8A) and male (Fig. 8B) mice. Female NMS-HHAID mice had significantly lower serum corticosterone levels compared to female naïve-control mice (Fig. 8A), whereas male NMS-HHAID mice had significantly higher serum corticosterone levels compared to NMS-control and NMS-HHD mice (Fig. 8B). Male NMS-HHD mice also had significantly lower serum corticosterone levels compared to naïve-HHD mice. These results suggest that there is a profound sexual divergence in diet impact on corticosterone levels, particularly in response to HHAID.

Fig. 8.

Effects of diet and early life stress on serum corticosterone concentrations. (A) In female mice, there was a significant effect of diet (p = 0.008) on serum corticosterone level. NMS-HHAID had significantly lower serum corticosterone compared to NMS-control mice. (B) In male mice, there was also a significant effect of diet (p = 0.021) on serum corticosterone levels. NMS-HHD mice had significantly lower serum corticosterone compared to naïve-HHD and NMS-HHAID mice and NMS-HHAID mice also had significantly higher corticosterone compared to NMS-control. δ denotes a significant effect of diet, two-way ANOVA. *p < 0.05 vs. same diet-fed naïve, &, &&&p < 0.05, 0.001 HHAID vs. control, ##p < 0.01 HHAID vs. HHD, Fisher’s LSD posttest. n = 6–16 per group.

Diet and NMS significantly increase mRNA levels of macrophage and inflammatory markers in periovarian and epididymal adipose tissue

In female mice, there was a significant effect of diet on the mRNA levels of F4/80 (general macrophage marker), CD68 (general macrophage marker), CD11b and CD11c (anti- and pro-inflammatory macrophage markers, respectively), and TNFα (pro-inflammatory cytokine) (Fig. 9A). For nearly each gene, female HHAID-fed mice had significantly higher mRNA levels compared to control- and HHD-fed mice. In males, there was a significant effect of diet on CD68, CD11c, and IL-10 (anti-inflammatory cytokine) mRNA levels (Fig. 9B).

Fig. 9.

The effect of diet and early life stress on gene expression of inflammatory markers in gonadal adipose tissue. (A) In females, there was a significant effect of diet on F4/80 (p < 0.0001), CD68 (p < 0.0001), CD11b (p = 0.001), CD11c (p = 0.024), and TNFα (p = 0.016). HHAID-fed mice had significantly higher mRNA levels compared to both control-fed and HHD-fed for most genes analyzed, regardless of NMS status. (B) In males, there was a significant effect of diet on CD68 (p = 0.003), CD11c (p < 0.0001), and IL-10 (p = 0.021) and additional effects of NMS (p < 0.0001) and an interaction effect of NMS and diet (p = 0.012) on CD11c. F4/80 mRNA levels were significantly higher in NMS-HHAID mice compared to NMS-control or NMS-HHD mice. CD11c was significantly higher in NMS-HHAID and -HHD mice compared to NMS-control mice and to naïve mice fed the same diet. Il-10 was elevated in naïve-HHAID mice compared to naïve-Control mice. δ denotes a significant effect of diet, two-way ANOVA. *p < 0.05 vs. same diet-fed naïve, †p < 0.05 HHD vs. control, &p < 0.05 HHAID vs. control, #p < 0.05 HHAID vs. HHD, Fisher’s LSD posttest. n = 5 per group.

There was also a significant effect of NMS on CD11c mRNA levels. Male NMS-HHAID and -HHD-fed mice had significantly higher CD11c mRNA levels compared to NMS-control-fed mice and male naïve-HHAID-fed mice had a significantly higher IL-10 mRNA level compared to naïve-control-fed mice. These results suggest that HHAID largely drives periovarian adipose inflammatory marker expression in female mice, regardless of NMS status, whereas diet negatively impacts inflammatory markers only in male NMS mice.

DISCUSSION

Chronic pain is a highly prevalent and costly condition experienced by millions of Americans (Dahlhamer et al., 2018). It is associated with an increased inflammatory state and comorbid expression of other chronic pain disorders (Aaron and Buchwald, 2001; Arnold et al., 2006; Clemens et al., 2008a, 2008b; Rodriguez et al., 2009). In the present study, we used a model of NMS in mice to test whether a HHD or HHAID had an effect on early life stress-induced widespread hypersensitivity (Pierce et al., 2014, 2016; Fuentes et al., 2015, 2017; Fuentes and Christianson, 2018). We hypothesized that mice consuming a HHD would be protected from developing NMS-induced urogenital and hindpaw hypersensitivity, compared to a low-fat control diet and we further hypothesized that the HHAID would confer additional benefits. Our results demonstrate a benefit of only HHAID on most pain-like behavioral measurements, yet highlight the complex relationships between diet, stress, and sex on body weight and metabolic regulation, markers of inflammation, and corticosterone concentrations. Interestingly, the HHAID improved measures of hypersensitivity despite inducing increased adiposity and evidence of metabolic dysfunction, particularly in female mice, highlighting an important disconnect between the comorbid and highly prevalent chronic conditions of pain and obesity.

Consumption of the HHAID increased perigenital withdrawal thresholds in all groups regardless of sex or arly-life stress exposure. These findings are in line with studies which found improvements in pain outcomes following administration of individual anti-inflammatory compounds (Nah et al., 2000; Sharma et al., 2007; Kuang et al., 2012; Di et al., 2014; Kuptniratsaikul et al., 2014; Wang and Wang, 2017; Silva et al., 2019). Many of these studies (Nah et al., 2000; Sharma et al., 2007; Kuang et al., 2012; Wang and Wang, 2017) were conducted in naïve, male mice with quantities of anti-inflammatory compounds that were 2–1000 times the amount used in the current study. Our study was unique in that it included both naïve and stressed conditions, female and male mice, and quantities of anti-inflammatory compounds that are realistically attainable in the human diet (Totsch et al., 2015). There is evidence that this type of dietary intervention may translate well to humans as a recent systematic review (Brain et al., 2019) found that nutritional interventions significantly reduced pain scores, although the specific interventions were varied and additional high-quality clinical trials are needed. The inclusion of both sexes in this study led to some interesting albeit expected sex differences.

We found that only the female mice displayed NMS-induced hindpaw hypersensitivity suggesting that female mice are more susceptible to widespread sensitivity than male mice. It is widely accepted that this sex difference exists in the incidence of chronic pain disorders as well as in laboratory-evoked pain responses (Sorge and Totsch, 2017). Females are over represented in most pain conditions that present in both sexes including fibromyalgia, migraine, temporomandibular disorder, irritable bowel syndrome, and interstitial cystitis/painful bladder syndrome (Berkley, 1997). Additionally, females exhibit increased sensitivity to evoked painful stimuli compared to males (Fillingim, 2000; Mogil, 2012). The mechanisms behind the sex differences in pain are attributed to a complex interaction between biological and psychosocial mechanisms. Preclinical work has revealed that pain processing is a complex circuitry that involves the peripheral and central nervous systems and both inhibitory and facilitatory mechanisms and sex hormones are known to influence this pathway at multiple levels through complicated interactions (Fillingim and Ness, 2000). This sex difference in hindpaw hypersensitivity is in line with previous findings and highlights the importance of studying both sexes in the field of pain research.

The HHD decreased perigenital withdrawal thresholds in NMS male, but not female, mice. We hypothesize this was due to the high fat content (35%), similar to the increase in hindpaw allodynia that was previously found in males on 54% high fat diet (Cooper et al., 2017). The anti-inflammatory components in the HHAID prevented this high fat diet-induced hypersensitivity, independent of body weight. These findings indicate that while females may generally be more susceptible to widespread sensitivity, the addition of high fat diet can reverse this sex difference and confer a greater susceptibility to allodynia in males, which coincides with the males increased risk of developing high fat diet-induced adverse metabolic effects.

Patients with chronic pain syndromes often experience symptoms of or are diagnosed with mood disorders, such as anxiety and depression (Demyttenaere et al., 2007; Clemens et al., 2008a, 2008b; Gureje et al., 2008). Anhedonia, ora lack of pleasure-seeking behavior, is a hallmark symptom of depression and can be measured behaviorally in rodents (Schmidt et al., 2011). Nest building is an innate and complex behavior that is impacted by hippocampal damage (Deacon et al., 2002) and social defeat stress (Otabi et al., 2017). The latter impact can be reversed by treatment with antidepressants, suggesting that nest building may be an appropriate measure for evaluating depressive-like behaviors. Here, we saw an overall diet effect on lowering nest building scores in the female, but not male, mice. Specifically, the female NMS mice on the HHAID showed a significantly lower nest score compared to naïve control-fed mice. We have previously reported on reduced regulatory gene expression in the hippocampus of NMS mice (Pierce et al., 2014, 2016, 2018; Fuentes et al., 2017). These results suggest that HHAID may be exacerbating these deficits, thereby inducing a depressive-like state in these mice. Future studies are needed to fully explore this potential outcome.

It is well-known that males, compared to females, are more prone to the adverse health effects of high fat diet (Hwang et al., 2010; Gelineau et al., 2017; Foright et al., 2020). Similar to our previous work using a 45% high fat diet (Eller et al., 2020), all male mice gained substantially more body weight and fat mass on the 35% fat diets (HHD and HHAID) compared to the 10% fat control diet. This was largely driven by increased caloric intake, although changes in energy expenditure and ambulatory cage activity were not measured and may also contribute to the phenotype. In line with this increased adiposity and body weight, markers of metabolic health including glucose intolerance and, in some cases, fasting insulin, were worsened by the HHAID and HHD in male mice. Future work can determine whether these changes in glucose tolerance were the result of altered glucose clearance or insulin secretion.

An elevation in caloric intake was also seen in female mice fed either the HHAID or HHD, however, it was only the NMS-HHAID female mice that displayed elevated body weight and adiposity. This suggests that energy expenditure was affected by the added anti inflammatory components, despite the negligible calorie content, and enhanced weight and adiposity gains following early life stress. This was surprising because the HHAID and HHD were derived from identical macronutrient sources/composition and energy density. It is also interesting that improvements in pain-like behaviors did not reduce weight gain or adiposity in NMS-HHAID-fed mice. This is similar to what was found in a different study using this HHAID, in which the HHAID-induced prevention of mechanical and thermal sensitivity in a model of inflammatory pain did not prevent increases in body weight and adiposity (Totsch et al., 2018) or an exercise intervention that reversed high fat diet-induced mechanical allodynia but did not alter body weight (Cooper et al., 2017). There is a positive association between pain and body mass index (Stone and Broderick, 2012) that is often assumed to be causally related. One prevailing hypothesis is that high levels of pain discourage physical activity, resulting in weight gain due to insufficient energy expenditure relative to energy intake. These data do not support this hypothesis but instead suggest that conditions that drive the development of chronic pain, such as NMS used in this study, may lead to long term adaptations in systems that regulate energy intake and expenditure in a manner that is independent of the development of long-term pain-like behaviors. It is worth noting that the circuitry regulating acute states of hunger and pain, overlap in complex ways that are only just beginning to be understood and may play a role in this model (Alhadeff et al., 2018).

We have previously found altered corticosterone concentrations, a marker of HPA axis output, in this model of early life stress (Fuentes et al., 2017; Pierce et al., 2018). This is consistent with the clinical literature which shows both hypercortisolism (Heim et al., 2001; Tyrka et al., 2008) and hypocortisolism (Heim et al., 1998; Gunnar and Quevedo, 2008) in adults that report a history of childhood abuse or stress. We did not find a main effect of NMS in this cohort, although the measurement at the single timepoint (early in the light cycle) may have masked disruptions in the cyclicity of the circadian secretions or differences in peak concentrations. We did find a sex-specific main effect of diet in which the female mice on either of the diets high in fat (HHAID and HHD) displayed reduced corticosterone concentrations compared to control diet-fed females. The reason for this decrease is unclear, although it is known that high fat diet can decrease adipose tissue 11 β-hydroxysteroid dehydrogenase type 1 (11β-HSD1) (Morton et al., 2004; Drake et al., 2005), which converts inactive corticosterone into active corticosterone (Tomlinson et al., 2004). Additionally, its known that high fat diet, through endogenous opioid secretion (Tsujii et al., 1987a, 1987b), can dampen HPA axis activation (Drolet et al., 2001). Both of these mechanisms would theoretically decrease the amount of circulating corticosterone; however, future studies are needed to determine the contribution of either of these mechanisms to this phenotype. It is unclear why the females consistently saw this diet effect while the males did not, although sex differences in the HPA axis including corticosterone secretion and availability are widely known and influenced by the sex hormones (Wigger and Neumann, 1999; Kalinichev et al., 2002; Kudielka and Kirschbaum, 2005; McCormick et al., 2005).

Manv chronic nain disorders are associated with a chronic state of low-grade inflammation (Zhang and An, 2007; Scarpellini and Tack, 2012; Miller and Raison, 2016) and diets similar to the HHAID have been shown to attenuate or prevent the development and pain in humans (Keske et al., 2015) and rodents (Bose et al., 2008; Weisberg et al., 2008; Jeon et al., 2012; Choi et al., 2014). We expected the HHAID-induced improvements in pain-like behaviors to coincide with reductions in tissue inflammation. Not only did the gene expression profile of pro-inflammatory markers within the gonadal fat pad not show reductions but, in some cases, levels were elevated in the NMS mice (F4/80 (females) and CD68 (males and females)). These findings may have been complicated by the increased adiposity and visceral fat of the NMS-HHAID-fed mice as visceral adipose accumulation is associated with chronic low-grade inflammation (Weisberg et al., 2003; Lumeng et al., 2007; Gregor and Hotamisligil, 2011). Although we did not measure inflammation in other peripheral or central tissues, it is possible there were improvements in inflammation in those tissues that contributed to improvements in the widespread hypersensitivity independent of the inflammation in this adipose tissue depot. Future studies could benefit from preventing the increased adiposity by restricting intake on the HHAID or the addition of a low-fat diet supplemented with the anti-inflammatory compounds to determine how an HHAID affects inflammation in the absence of increased adiposity.

In summary, the HHD had no effect on hindpaw withdrawal thresholds and even worsened perigenital withdrawal threshold in NMS male mice. In contrast, the HHAID improved perigenital withdrawal thresholds in all mice regardless of sex or early life stress exposure, and selectively improved hindpaw allodynia in female NMS mice. This improvement in pain-like behaviors was seen despite HHAID-induced increases in body weight and adiposity, glucose intolerance, and pro-inflammatory markers. Both the HHAID and HHD increased body weight and adiposity compared to the control diet suggesting that, at least in this preclinical model, the high fat diet-induced increase in food intake was metabolically detrimental and was not overcome by the inclusion of anti-inflammatory compounds which, in the case of the NMS females, worsened weight gain and adiposity. If these findings translate to clinical populations, it may suggest that there are limits to the overall benefit of an anti-inflammatory diet if total intake is not limited.

Abbreviations:

AUC

area under the curve

CCI

chronic constriction, injury

EGCG

epigallocatechin gallate

HHAID

healthy human anti-inflammatory, diet

HHD

healthy human diet

HPA

hypothalamic–pituitary–adrenal

NMS

neonatal maternal separation

GTT

glucose, tolerance test