Activation of the heat shock response by human milk-derived extracellular vesicles in neonates with perinatal high-fat diet exposure

Abstract

Maternal consumption of a high-fat diet (mHFD) during perinatal life influences hypothalamic-pituitary-adrenal (HPA) axis activation and impacts the long-term physiological and metabolic health of offspring. Milk-derived extracellular vesicles (MEVs) are lipid-coated nanovesicles that transfer biological materials from mother to infant and can survive intestinal degradation and cross the blood-brain barrier. MEVs provide cytoprotection in peripheral organs; however, their pro-survival functions remain unknown in the neonatal brain. Further, sex differences resulting from MEV treatment require investigation, as male and female neonates display variable responses to early life nutrient stress. We investigated the interaction between MEVs and the heat shock protein response in the liver, hypothalamus, and prefrontal cortex in male and female neonatal rats exposed to perinatal mHFD at postnatal day 11. MEV treatment robustly modulated the HSR in female neonates with the largest response recorded in the prefrontal cortex. These results suggest that MEVs may influence pro-survival outcomes in the prefrontal cortex by activating HSF1-mediated pro-survival in a sex-specific manner.

Introduction

Perinatal (the combined prenatal and postnatal periods) development is sensitive to maternal factors. ¹ Maternal malnutrition resulting from over- or under-consumption of macro/micronutrients influences developmental programming and impacts the long-term health of offspring. ^2–5 Specifically, the maternal consumption of a high saturated fat diet (mHFD) increases the risk of developing congenital defects, metabolic disorders, and neurodevelopmental and neuropsychiatric disorders in offspring. ^6–10 Studies conducted in murine and ovine models have demonstrated that mHFD exposure during the perinatal period alters metabolism; decreases brain weight, bone density, and muscle development; increases adiposity and bodyweight; and increases pro-inflammatory cytokine production and circulation. ^5,7,11–15 These exposures may also lead to sex-specific physiological and epigenetic changes in offspring that persist into adulthood. ^16–18

Perinatal exposure to mHFD also leads to the activation and chronic potentiation of the hypothalamic-pituitary-adrenal (HPA) axis. ^19–21 The HPA axis is a neuroendocrine system responsible for regulating glucocorticoid levels during stress. ²² Several organs are implicated in the HPA-mediated stress response. The hypothalamus is the main HPA axis regulator. ²³ During homeostasis, the hypothalamus regulates energy metabolism, thermoregulation, circadian cycles, and feeding behavior. ^24,25 During stress, however, the paraventricular nucleus (PVN) of the hypothalamus activates the HPA-mediated stress response through the release of corticotrophin releasing hormone (CRH) and arginine vasopressin (AVP). ^26,27 CRH and AVP stimulate the anterior pituitary gland to release adrenocorticotrophic hormone (ACTH) into general circulation and induces the release of glucocorticoids from the adrenal glands. ^26,27 Lesion and glucocorticoid inhibition studies have elucidated the role of the prefrontal cortex as a site of glucocorticoid receptor-mediated feedback inhibition of the HPA axis. ^23,28–30 The liver is also impacted by HPA-mediated neuroinflammation via the liver-gut-brain axis. ³¹

The activation of the HPA axis and the induction of the heat shock protein response (HSR) are core biological processes that help maintain cellular and organismal homeostasis. Heat shock proteins (HSPs) are molecular chaperones that participate in the initiation of protein folding, refolding, repair, disaggregation, and degradation. ³² In addition to their homeostatic functions, HSPs are upregulated in response to environmental stressors including oxidation and heat and cold shock. ^33–35 In a study by Blake et al. ³⁶ restrain stress in rats was shown to activate the HPA axis and the expression of candidate HSPs. Interestingly, mHFD-induced metabolic dysfunction is also associated with HSR dysregulation. ^34,37,38

Mammalian milk is a complex, heterogeneous biological fluid tailored to meet the energetic and developmental requirements of newborns. ^3,39 Milk composition is sensitive to maternal diet. Despite studies illustrating mHFD-induced changes in select milk components, maternal milk feeding is beneficial and necessary to support the healthy development of offspring as the transmission of a plethora of biologically active molecules in milk mitigate the negative developmental outcomes associated with gestational mHFD. ^{2–4,40–43}

One type of biologically active molecule in mammalian milk is milk-derived extracellular vesicles (MEVs). MEVs are lipid-coated nanovesicles (30–150 nm in size) with immunomodulatory and anti-inflammatory properties. ^44,45 MEVs have been reported to decrease rates of apoptosis, attenuate NFκB-mediated pro-inflammation, and reduce the progression of gastrointestinal diseases, including necrotizing enterocolitis and inflammatory bowel disease, in neonates across experimental models. ^46–53

Despite the known anti-inflammatory and cytoprotective potential of MEVs in peripheral organs, the ability of MEVs to interact with and enhance existing pro-survival responses in the neonatal brain require further investigation. The main objective of our study is to explore the association between MEVs and the HSR in the liver, hypothalamus, and prefrontal cortex of offspring with mHFD exposure. To our knowledge, our study is the first to explore the relationship between MEVs and the HSR in male and female neonatal rats in the context of promoting cytoprotection in HPA-axis associated tissues.

Materials and methods

MEV isolation and characterization

Unpasteurized human donor milk (n = 11 anonymous donors) was obtained from NorthernStar Mothers Milk Bank (Calgary, AB, Canada) and pooled into a homogeneous mixture to limit variability in milk composition between the donors. Milk was processed and MEVs were isolated using serial ultracentrifugation and filtration as per Wijenayake et al. ⁴¹

MEVs were characterized in accordance with the Minimum Information for Studies of Extracellular Vesicles (MISEV) 2023 guidelines ⁵⁴ and as described by Wijenayake et al. ⁴¹ Briefly, MEV particle size and concentration were quantified using nanoparticle tracking analysis (Malvern Instruments Ltd.; NanoSight NS300) at the Structural and Biophysical Core Facility at the Hospital for Sick Children (Toronto, ON, Canada). MEV morphology and integrity was characterized using transmission electron microscopy (FEI Talos F200x S/TEM) at the Manitoba Institute of Materials (Winnipeg, MB, Canada). MEV biomarkers were characterized using western immunoblotting against CD9 and CD81, and syntenin-1. The results of the MEV characterization are available in Storm et al. ⁵⁵ The same batch of MEVs isolated from the same 11 donors were used across studies.

Animal care and treatment

All experimental procedures were approved by the University of Winnipeg Animal Care Committee (AE18072) and were in accordance with the guidelines of the Canadian Council on Animal Care. 7-week-old female (n = 12) and male Long Evans rats (n = 6) (strain code: 006) were purchased from Charles River Canada (St. Constant, QC). All rats were housed in same-sex pairs until mating and maintained on a 12h:12h light/dark cycle with ad libitum access to food and water. Following one-week of acclimatization, a subset of females (n = 6) was placed on a mCHD consisting of 10% kcal fat (Research Diet Inc., D12450J), with the remaining females (n = 6) placed on a mHFD consisting of 60% kcal fat (Research Diet Inc., D12492) (Fig. 1). The mCHD was matched in sucrose content to the mHFD. Diet was maintained for 4 weeks prior to mating, and throughout gestation and lactation. ^2,5,17,56 For breeding, females were pair housed with one male for seven days, after which females were housed individually throughout gestation and lactation. Pregnancy was inferred by monitoring daily body weight gain rather than using vaginal smears, allowing for reduced manipulation and maternal stress for dams. Vaginal plugs were not used because they do not guarantee pregnancy. ⁵⁷

Figure 1. Animal care and study overview. Following one week of acclimatization, dams were placed on a control diet (mCHD) consisting of 10% kcal fat, or a high-fat diet (mHFD) consisting of 60% kcal fat, (n = 6/diet). Diet was maintained for four weeks prior to mating, throughout mating and gestation, and lactation. After parturition, at postnatal day (PND) 2, litters were weighed and culled to 12 pups/litter (n = 6 females and n = 6 males, where possible) to standardize maternal care provisions across litters. MEV treatment began at PND4, where a subset of neonates were controls (mCHD or mHFD) or received a vehicle gavage (mCHD-PBS or mHFD-PBS) or received MEV gavage (mCHD-MEV or mHFD-MEV) (n = 1–2 pups/litter/sex). Oral gavage (1 × 10¹⁰ MEVs/g of body weight) was administered twice a day, 6h apart, until PND11. At euthanasia on PND11, liver, hypothalamus, prefrontal cortex, stomach milk curd, and retroperitoneal fat were collected for downstream molecular analysis.

Dam weight gain and caloric intake was measured once a week prior to mating, and throughout gestation, then daily during lactation. Day of parturition is considered PND0 and the day before is considered the last day of gestation. At PND2 litters were weighed and culled to 12 neonates/litter (n = 6 females and n = 6 males, where possible) to standardize maternal care provisions across litters. Maternal care differences resulting from consuming this specific HFD is reported to be negligible in Long Evans dams. ^57,58

MEV treatment via oral gavage began at PND4 and continued until the end of the study at PND11. MEVs were administered using disposable 22-gauge polypropylene feeding tubes (Instech Laboratories: FTP-22-25). n = 1–2 neonates/litter/sex were randomly assigned to three treatment groups, depending on sex ratio and litter sizes: (1) neonates that were separated from the nest with their littermates and handled but did not receive gavage (referred to as mCHD or mHFD), (2) neonates that received the vehicle control of 1X PBS (referred to as mCHD-PBS or mHFD-PBS), and (3) neonates that received 1 × 10¹⁰ MEVs in 1X PBS per gram of bodyweight (referred to as mCHD-MEV or mHFD-MEV). Oral gavage was administered twice a day, during the light phase, 6h apart. The oral gavage dosage used in our study is similar to previous studies. ^59–61 MEVs have a high cross-species tolerance and the administration of human and/or bovine MEVs to murine species does not affect viability or induce physiological effects. ^59,62,63 Due to the large volumes of rodent milk that is required to isolate sufficient MEVs to gavage neonates twice a day, for 7 days, it was not feasible to use Long Evans rat milk for this experiment. The neonates were separated from the nest for a maximum of 15 minutes/procedure and 30 minutes/day. As per previously published works, an elevation in plasma corticosterone (ng/ml) was not observed until >2h maternal separation. ^64–66 Pup body weight and naso-anal length were determined daily from PND2 to PND11. The Lee index is used as an index of obesity in rodents and has been shown to correlate to carcass fat percentage. ⁶⁷ The Lee index was calculated as follows as per Simson and Gold ⁶⁷ :

$${\rm{Lee}}\;{\rm{index}}\; = {{\root 3 \of {{\rm{body}}\;{\rm{weight}}\;\left( {\rm{g}} \right)} } \over {{\rm{naso - anal}}\;{\rm{length}}\;\left( {{\rm{cm}}} \right)}} \times 1000$$

At PND11 neonates were euthanized via swift decapitation. The liver, hypothalamus, prefrontal cortex, stomach milk curd, and retroperitoneal fat were collected, flash frozen, and stored at −80 °C. In total, 32 female and male neonates at PND11 from n = 8 litters were used for this study. n = 2 females and n = 2 males were used per litter with n = 4 animals per diet per treatment per sex.

The following resource equation described in Arifin et al. ⁶⁸ for preclinical experimental rodent studies comparing groups was used to calculate the minimum and maximum sample sizes that are required for molecular analysis.

$$ N=\left({10 \over k}\right)+1 $$

N = Sample size

k (Number of groups) = 8 (2 diets × 2 treatments × 2 sex)

Minimum N = (10 / 8) + 1 = 2.25, rounded up to 3 animals/group, total sample size = 24

Maximum N = (20 / 8) + 1 = 3.5 – rounded up to 4 animals/group, total sample size = 32

RNA extraction and cDNA synthesis

Total soluble RNA (≥ 18 nucleotides) was extracted from approximately 30 µg of frozen liver, hypothalamus, and prefrontal cortex (n = 4 biological replicates/diet/treatment/sex) using TRIzol reagent (ThermoFisher Scientific: 15596018) according to the manufacturer’s instructions. RNA samples were resolved on a 1% TAE-agarose gel with 2x RNA loading dye (1:1, v/v) (Life Technologies: R0641) stained with Red Safe dye (FroggaBio: 21141) to verify stability and integrity (Supplementary Figure S1–S3 ). RNA concentration (ng/µL) and purity (A260:280 and A260:230 ratios) were determined using a Nanodrop One/One^C Microvolume-UV/Vis spectrophotometer (ThermoFisher Scientific: ND-ONE-W). RNA concentrations for each tissue are listed in Supplementary Table S1. Genomic DNA was removed using DNase I treatment, and a RNA Clean & Concentrator kit (Zymo Research: R1017) was used to purify select RNA samples with an A260:A280 ratio <1.8, according to the manufacturer’s instructions. Samples with an A260:280 ratio of 1.8–2.0 were used for cDNA synthesis. 2000 ng of total soluble RNA was reverse transcribed into cDNA using the High-Capacity cDNA Reverse Transcription kit (Applied Biosystems: 4368814) according to the manufacturer’s instructions. cDNA was synthesized using a T100™ Thermal Cycler (Bio-Rad: 1861096) (amplification parameters: 25 °C for 10 mins; 37 °C for 120 mins; 85 °C for 5 mins; hold at 4 °C). Samples were stored at −80 °C for future use.

Primer design and RT-qPCR

Gene expression of HSF1, HSPA1B, HSP90AA1, and DNAJB1 were measured via RT-qPCR using a QuantStudio™ 5 PCR system (Applied Biosystems: 96-well and 0.2 mL block) using Fast SYBR™ Green Master Mix chemistry (Applied Biosystems: 4385612). Internal controls were GAPDH and YWAZ (liver), GUSB (hypothalamus), and 18S rRNA and YWAZ (prefrontal cortex). Additional candidate internal controls were tested in the liver, hypothalamus, and prefrontal cortex but were determined unsuitable for normalization because they varied across treatment or diet in either sex (Supplementary Table S2). Primers were designed using nucleotide sequence information from NCBI or obtained from previous literature. Sequence information from NCBI and the OligoAnalyzer™ Tool from Integrated DNA Technologies (IDT, Coralville, Iowa, USA) were used to assess primer compliance with the Minimum Information for Publication of Quantitative Real-time PCR Experiments (MIQE) guidelines. ⁶⁹ Primer parameters are listed in Supplementary Table S3, and primer bioinformatics are listed in Supplementary Table S4 .

Annealing temperatures of primer pairs were determined by testing a range of temperatures ± 5 °C of the forward primer (5′–3′) melting temperature using the QuantStudio™ 5 PCR system Veriflex settings. Temperature testing was conducted using a pool sample (10 ng/µL) representing all test samples across tissues. Melt curve analysis was conducted to determine primer pair specificity. Primer pairs that generated a single, sharp peak with a derivative reporter >200,000 and devoid of primer dimers were used for quantification. Optimal cDNA loading amounts were determined per primer pair and per tissue using an 8-point standard curve ranging from 500 ng/µL to 3.91 ng/µL. Analyses were conducted in triplicate and an inter-plate converter was used across plates to account for variability between plates and amplifications.

Protein extraction and analysis of abundance by western immunoblotting

Total soluble protein (n = 4 animals/diet/treatment/sex) was extracted according to Tessier et al. ⁷⁰ Animals used in the protein analysis are littermates of the animals that were used in the RNA analysis. Protein concentration was determined using a BCA assay (Pierce™ BCA Protein assay; ThermoFisher Scientific: 23227), according to the manufacturer’s instructions, and absorbance readings were obtained using a BioTek Synergy H1 Multimode microplate reader (Agilent: BTSH1M2SI) at 562 nm. Lysates were combined with bromophenol blue loading dye with SDS (100 mM Tris-base; 4% w/v, SDS; 20% v/v, glycerol; 0.2% w/v, bromophenol blue) (Bioshop: SDS001.1) and β-mercaptoethanol (10%, v/v) (Bioshop: MERC002.500), then vortexed and heated to 95 °C for 10 minutes for denaturation. Samples were stored at -20 °C for future use.

Protein lysates were resolved using 10%–15% SDS-polyacrylamide gels. A standard curve (5 μg–35 μg) using an organ-specific pool sample was used to determine optimal protein amount to load per antibody and per tissue. PageRuler™ Plus prestained protein ladder (ThermoFisher Scientific: 26619) was used as a molecular weight standard, and a liver pool was run in duplicates as the interblot converter to be used for normalization and comparisons across immunoblots. Samples were resolved (n = 4/treatment group/sex/diet) on SDS-polyacrylamide gels for 110 minutes in 1x tris-glycine running buffer (0.3%, w/v Tris-base; 14.4%, w/v glycine; 1%, w/v SDS) at 180 V in a Sub-Cell GT electrophoresis cell (Bio-Rad: 1704401). Proteins were transferred onto 0.45 µm PVDF membranes (Bio-Rad: 1620174) using a Trans-Blot Turbo Transfer System (Bio-Rad: 1704150). Membranes were blocked using 1%–20% casein-TBST solution (30 minutes, 22 °C), incubated with primary antibody (1:500 or 1:1000, v/v), followed by goat HRP-conjugated anti-rabbit IgG secondary antibody (1:10,000 or 1:15,000, v/v) for 45 minutes at 22 °C. The immunoblots were visualized using western ECL substrate solution (ThermoFisher Scientific: 34580) and chemiluminescence imaging (ChemiDoc™ MP imaging system; Bio-Rad: 12003154) with Image Lab Software (version 6.1). The immunoblots were stained for 2 minutes in Coomassie brilliant blue (BioShop: CBB555.10) solution (0.25%, w/v Coomassie blue salt; 7.5% acetic acid; 50%, v/v methanol) and destained for 5 minutes with destain solution (25%, v/v methanol; 10%, v/v acetic acid) at room temperature, on a rocker. ImageJ (version 1.5.3) was used to quantify protein abundance as per Abràmoff et al. ⁷¹ Western immunoblotting parameters for HSR targets are listed in Supplementary Table S5 and antibody information in Supplementary Table S6 . ECL and Coomassie images for HSR targets are in Supplementary Figure S4–S7 .

Statistical analysis

Statistical analysis was carried out using SPSS version 29.0.2.0 (IBM Corp.) and figures were constructed using GraphPad Prism 7 and BioRender.

Normality was assessed using Shapiro–Wilk test (p > 0.05). Equality of variance was tested with Levene’s Test (p > 0.05). All biological samples were independent by design. Extreme outliers with an IQR > 3 were identified using the SPSS boxplot outlier function and removed from the dataset, when necessary. Data that did not achieve normality were analyzed with Mann–Whitney U (for 2 comparisons) and Kruskal–Wallis H tests followed by Dunn’s post hoc tests (for 3 or more comparisons).

Factorial (time, diet [mCHD or mHFD], treatment [control, PBS, MEV], sex) repeated-measures general linear model (GLM) was used with Bonferroni correction for maternal bodyweight and average caloric intake during pre-gestation, gestation, and lactation, and offspring bodyweight and Lee index. Greenhouse-Geisser corrections were used as the datasets failed Mauchly’s tests of sphericity. Dam mating success, litter size, and postnatal weight gain were determined using two-tailed, independent samples t-tests to compare mCHD and mHFD litters.

Univariate GLM (diet, treatment, sex) was conducted to analyze total retroperitoneal fat weight and stomach curd weight, as well as for parametric transcript and protein abundance data in offspring. Tukey HSD post hoc test was used to conduct pairwise comparisons between treatment groups. A Pearson correlation was used to compare offspring bodyweight (g) and retroperitoneal fat weight (g), and offspring bodyweight and stomach milk curd weight (g).

All transcript and protein abundance data were analyzed with sexes separated, as perinatal mHFD exposure leads to sex differences. ^5,17,57 n = 2 neonates per litter per sex were used for transcript and protein analyses, as such dam ID was not used as a covariate. The core results of the study compare male and female offspring across two diets (mCHD and mHFD) and two treatments (controls and MEV gavaged) for n = 4 animals/diet/treatment/sex. The raw data for neonates who received the 1X PBS vehicle is in Supplementary Table S7 . A p-value < 0.05 was considered statistically significant. Data are represented as mean ± SEM.

Results

Changes in dam bodyweight, caloric intake, mating success, litter size, and postnatal weight change

Dam bodyweight remained unchanged throughout pre-gestation, gestation, and lactation between mCHD and mHFD (Fig. 2a). Specifically, there were no significant differences in bodyweight between mCHD and mHFD dams (F(_1,9) = 0.230, p = 0.643), nor significant interactions between the diet conditions and weeks spent on respective diets (F(_{1.882, 16.941}) = 0.357, p = 0.693). Dam body weight increased with age in mCHD and mHFD (F(_{1.882, 16.941}) = 82.507, p < 0.001). Similarly, there were no significant changes between mCHD and mHFD dams in mating success (66.7% success for mCHD and mHFD), sizes of litters (t(8) = −1.452, p = 0.185), or postnatal weight change (t(8) = −1.549, p = 0.160) (Fig. 2b–2d). Average daily caloric intake of dams was analyzed throughout pre-gestation, gestation, and lactation (Fig. 2e). mHFD dams had a higher average daily caloric intake than mCHD dams (F(_1,9) = 13.079, p = 0.006), where mHFD dams consumed more kCal per day in the first and second weeks of pre-gestation. Caloric intake also increased over time in both mCHD and mHFD dams (F(_{1.313, 11.820}) = 32.726, p < 0.001), although there was no significant interaction between diet conditions and weeks spent on the respective diets (F(_{1.313, 11.820}) = 2.075, p = 0.175). Daily caloric intake during the lactational period was also measured in dams post-parturition (Fig. 2f). During lactation, there are no significant differences in daily caloric intake between mCHD and mHFD dams (F(_{1, 8}) = 2.674, p = 0.141); however, as seen previously, there is an increase in caloric intake over time (F(_{1.921, 15.364}) = 3.835, p = 0.046). No significant interactions between diet conditions and days spent on the diet during lactation were observed (F(_{1.921, 15.364}) = 1.749, p = 0.207).

Figure 2. Maternal responses to high-fat diet (mHFD) consumption. (a) Changes in dam bodyweight during pre-gestation, gestation, and lactation, between control diet (mCHD) and mHFD (n = 6/diet). (b) Mating success of mCHD and mHFD dams. (c) Litter sizes at parturition. (d) Dam postnatal weight change between parturition and the end of the study at postnatal day (PND) 11. (e) Average daily caloric intake (kCal/day) in dams during pre-gestation, gestation, and lactation. (f) Daily caloric intake (kCal/day) in dams throughout lactation. *Main effect of diet (p < 0.05), **Main effect of time (p < 0.05), and #Interaction between diet × time (p < 0.05). Data presented are means ± standard error.

Change in offspring bodyweight and Lee index

Changes in offspring bodyweight were analyzed from PND2 until euthanized at PND11 (Fig. 3a). No differences in bodyweight were observed between male and female neonates within a diet (F(_{1, 98}) = 2.835, p = 0.095) or between treatment groups (F(_{2, 98}) = 0.240, p = 0.787); therefore, changes in offspring bodyweight were determined without separating neonates by sex or treatment. Exposure to mHFD influenced offspring weight and caloric intake. Offspring bodyweight increased throughout the lactational period (F(_{1.369, 142.428}) = 2086.469, p < 0.001) with significantly higher bodyweight in neonates borne to mHFD dams compared to mCHD dams (F(_{1, 104}) = 9.962, p = 0.002). Furthermore, a significant interaction between days and diet was observed (F(_{1.369, 142.428}) = 35.396, p < 0.001). No significant interactions were observed between days and treatment group (F(_{2.739, 142.428}) = 0.872, p = 0.449), nor days with diet and treatment (F(_{2.739, 142.428}) = 1.210, p = 0.307).

Figure 3. Offspring responses to maternal diet. (a) Changes in offspring bodyweight from postnatal day (PND) 2 to the end of the study at PND11 in response to control diet (mCHD) and mHFD (n = 24-31/diet/sex). (b) Changes in offspring Lee index throughout lactation between pups in mCHD and mHFD diet groups. *Main effect of diet (p < 0.05), **Main effect of time (p < 0.05), and #Interaction between diet × time (p < 0.05). Data were combined across sex and treatment as no main effects were seen. Data presented are means ± standard error.

Similar to bodyweight, changes in offspring Lee index were also analyzed from PND2 to PND11 (Fig. 3b). No differences in Lee index were observed between male and female neonates within a diet (F(_{1, 98}) = 0.008, p = 0.928) or between treatment groups (F(_{2, 98}) = 0.259, p = 0.773); therefore, changes in offspring Lee index were also determined without separating neonates by sex or treatment. Offspring Lee index decreased slightly throughout the lactational period (F(_{1.368, 142.265}) = 4.267, p = 0.029), although there were no significant differences between neonates borne to mCHD dams than mHFD dams (F(_{1, 104}) = 1.174, p = 0.281). Additionally, there were no significant interactions between days and diet (F(_{1.368, 142.265}) = 1.984, p = 0.155), days and treatment (F(_{2.736, 142.265}) = 0.998, p = 0.391), nor days with diet and treatment (F(_{2.736, 142.265}) = 1.036, p = 0.374).

Correlations between pup bodyweight, stomach milk curd weight, and retroperitoneal fat weight

At PND11, mHFD offspring had significantly higher total retroperitoneal fat weight than mCHD offspring (U = 1873.500, p = 0.002), however, there were no significant differences in fat deposition between males and females within a diet (U = 1599.500, p = 0.217) or between treatment groups (H(2) = 0.035, p = 0.982) (Fig. 4a). Therefore, changes were determined without separating neonates by sex or treatment. A positive linear correlation was observed between increased bodyweight and increased retroperitoneal fat weight at PND11 (R² = 0.3085, p < 0.0001) (Fig. 4b).

Figure 4. Analysis of offspring retroperitoneal fat and stomach milk curd weight in response to maternal diet. (a) Total retroperitoneal fat weight in neonates at postnatal day (PND) 11 in response to maternal control diet (mCHD) and maternal high-fat diet (mHFD). (b) Pearson correlation between offspring bodyweight and amount of retroperitoneal fat. (c) Total stomach milk curd weight in neonates at PND11 in response to mCHD and mHFD. (d) Pearson correlation between offspring bodyweight and amount of stomach milk curd present in stomach. *Main effect of diet (p < 0.05), **Main effect of time (p < 0.05), and #Interaction between diet × time (p < 0.05). Data were combined across sex and treatment as no main effects were seen. Data presented are means ± standard error.

mHFD offspring also had significantly higher stomach milk curd weight at PND11 when compared to mCHD offspring (U = 2069.000, p < 0.001), although there were no significant differences in stomach milk weight between males and females (U = 1532.000, p = 0.776) or between treatment groups (H(2) = 3.604, p = 0.165) (Fig. 4c). Therefore, changes were determined without separating neonates by sex or treatment. A positive linear correlation was also observed between increased bodyweight and increased stomach milk curd weight at PND11 (R² = 0.2221, p < 0.0001) (Fig. 4d).

Transcript abundance of HSR targets in offspring liver

In the liver, transcript abundance of HSF1 remained unchanged in male neonates (diet: (F(_1,11) = 1.993, p = 0.186); treatment: (F(_3,11) = 1.713, p = 0.222)) (Fig. 5a). In female neonates, while no diet effect was observed in HSF1 transcript (F(_1,12) = 0.007, p = 0.933), there was an effect of MEV treatment (F(_3,12) = 7.683, p = 0.004) with higher HSF1 transcript in mCHD-MEV neonates compared to mCHD (Tukey HSD, p = 0.006), lower HSF1 transcript in mHFD-MEV compared to mCHD-MEV (Tukey HSD, p = 0.034), and higher HSF1 transcript in mHFD compared to mCHD (Tukey HSD, p = 0.027) (Fig. 5a).

Figure 5. Transcript abundance of the heat shock response genes in the liver of male and female neonates at postnatal day (PND) 11 as determined by RT-(a) HSF1 liver transcript abundance. (b) HSPA1A liver transcript abundance. (c) HSP90AA1 liver transcript. (d) DNAJB1 liver transcript abundance. Quantity means are normalized to the geometric mean of two reference genes with stable expression: GAPDH and YWAZ. #Main effect of diet (p < 0.05). *Main effect of MEV treatment (p < 0.05). Pairwise comparisons between treatment groups are indicated with lowercase letters, where significant differences (p < 0.05) are denoted by different letters. mCHD: neonates born to mCHD dams that did not receive MEV supplementation. mCHD-MEV: neonates born to mCHD dams that received MEV supplementation. mHFD: neonates born to mHFD dams that did not receive MEV supplementation. mHFD-MEV: neonates born to mHFD dams that received MEV supplementation. n = 3-4 biological replicates/diet/treatment/sex. Data presented are means ± standard error.

HSPA1A transcript abundance remained unchanged in male neonates (diet: (F(_1,11) = 0.146, p = 0.710); treatment: (F(_3,11) = 0.961, p = 0.445)) (Fig. 5b). In female neonates, there was no effect of diet (F(_1,12) = 1.863, p = 0.197); however there was an effect of treatment (F(_3,12) = 7.808, p = 0.004) with higher HSPA1A transcript in mCHD-MEV neonates compared to mCHD (Tukey HSD, p = 0.003), lower HSPA1A transcript in mHFD-MEV compared to mCHD-MEV (Tukey HSD, p = 0.047) and lower HSPA1A transcript in mHFD compared to mCHD-MEV (Tukey HSD, p = 0.019) (Fig. 5b).

HSP90AA1 transcript abundance remained unchanged for male neonates (diet: (F(_1,12) = 1.966, p = 0.186); treatment: (F(_3,12) = 1.819, p = 0.197)) (Fig. 5c). In female neonates, there was no effect of diet (F(_1,12) = 1.696, p = 0.217) on HSP90AA1 transcript. A treatment effect was observed (F(_3,12) = 3.478, p = 0.050), although individual differences between treatments were not observed (Fig. 5c).

DNAJB1 transcript abundance remained unchanged in male neonates (diet: (F(_1,12) = 0.003, p = 0.956); treatment: (F(_3,12) = 0.011, p = 0.998)) (Fig. 5d). In female neonates, DNAJB1 transcript was responsive to changes in diet (F(_1,11) = 4.917, p = 0.049). There was also a treatment effect (F(_3,11) = 10.080, p = 0.002), with higher DNAJB1 transcript in mCHD-MEV compared to mCHD (Tukey HSD, p = 0.001), mHFD (Tukey HSD, p = 0.010) and mHFD-MEV (Tukey HSD, p = 0.006), respectively (Fig. 5d). Transcript data of male and female neonates that received PBS gavage (the sham controls) is included in Supplementary Table S7 .

Protein abundance of HSR targets in offspring liver

In the liver, HSF1 protein abundance in male neonates remained unchanged (diet: (F(_1,12) = 0.499, p = 0.494); treatment: (F(_3,12) = 2.538, p = 0.106)) (Fig. 6a). HSF1 protein abundance also remained unchanged in female neonates (diet: (F(_1,12) = 2.597, p = 0.133); treatment: (F(_3,12) = 2.965, p = 0.075)) (Fig. 6a).

Figure 6. Protein abundance of the heat shock response targets in the liver of male and female neonates at postnatal day (PND)11 as determined by western immunoblotting. (a) HSF1 liver protein abundance. (b) Hsp70 liver protein abundance. (c) Hsp90 liver protein abundance. (d) Hsp40 liver protein abundance. Protein targets are normalized to the abundance of total soluble proteins in the samples using Coomassie staining. ECL and Coomassie-stained images are displayed. #Main effect of diet (p < 0.05). *Main effect of treatment (p < 0.05). Pairwise comparisons between treatment groups are indicated with lowercase letters, where significant differences (p < 0.05) are denoted by different letters. mCHD: neonates born to mCHD dams that did not receive MEV supplementation. mCHD-MEV: neonates born to mCHD dams that received MEV supplementation. mHFD: neonates born to mHFD dams that did not receive MEV supplementation. mHFD-MEV: neonates born to mHFD dams that received MEV supplementation. n = 3-4 biological replicates/diet/treatment/sex. Data presented are means ± standard error.

Hsp70 protein abundance remained unchanged in male neonates (diet: (F(_1,12) = 0.227, p = 0.643); treatment: (F(_3,12) = 0.658, p = 0.594)) (Fig. 6b). Hsp70 protein abundance also remained unchanged in female neonates (diet: (F(_1,12) = 0.725, p = 0.411); treatment: (F(_3,12) = 1.181, p = 0.358)) (Fig. 6b).

Hsp90 protein abundance remained unchanged in male neonates (diet: (F(_1,12) = 0.064, p = 0.805); treatment: (F(_3,12) = 0.043, p = 0.987) (Fig. 6c). In female neonates, Hsp90 protein abundance was responsive to changes in diet (F(_1,12) = 10.665, p = 0.007). A treatment effect was observed (F(_3,12) = 3.561, p = 0.047), although no individual effects were observed between treatment groups (Fig. 6c).

Hsp40 protein abundance in male neonates was responsive to diet (F(_1,12) = 58.828, p < 0.001), and MEV treatment (F(_3,12) = 20.299, p < 0.001), where Hsp40 protein abundance was higher in mCHD than mHFD (Tukey HSD, p < 0.001) and mHFD-MEV (Tukey HSD, p = 0.004) (Fig. 6d). Similarly, Hsp40 protein abundance was higher in mCHD-MEV than mHFD (Tukey HSD, p < 0.001) and mHFD-MEV (Tukey HSD, p = 0.001). Hsp40 protein abundance in female neonates was also responsive to diet (F(_1,11) = 23.438, p < 0.001) and MEV treatment (F(_3,11) = 9.733, p = 0.002), where Hsp40 protein abundance was higher in mHFD-MEV than mCHD (Tukey HSD, p = 0.004) and mCHD-MEV (Tukey HSD, p = 0.004) (Fig. 6d). Protein data of male and female neonates that received PBS gavage (the sham controls) is included in Supplementary Table S7 .

Transcript abundance of HSR targets in offspring hypothalamus

In the hypothalamus, HSF1 transcript abundance remained unchanged in male neonates (diet: (F(_1,12) = 3.954, p = 0.070); treatment: (F(_3,12) = 1.908, p = 0.182)) (Fig. 7a). HSF1 transcript in female neonates was responsive to diet (F(_1,12) = 8.654, p = 0.012). The treatment effect approaches significance (F(_3,12) = 3.465, p = 0.051) (Fig. 7a).

Figure 7. Transcript abundance of the heat shock response genes in the hypothalamus of male and female neonates at postnatal (PND) 11 as determined by RT-qPCR. (a) HSF1 hypothalamus transcript abundance. (b) HSPA1A hypothalamus transcript abundance. (c) HSP90AA1 hypothalamus transcript abundance. (d) DNAJB1 hypothalamus transcript abundance. Quantity means are normalized to one reference gene with stable expression: GUSB. #Main effect of diet (p < 0.05). *Main effect of MEV treatment (p < 0.05). Pairwise comparisons between treatment groups are indicated with lowercase letters, where significant differences (p < 0.05) are denoted by different letters. mCHD: neonates born to mCHD dams that did not receive MEV supplementation. mCHD-MEV: neonates born to mCHD dams that received MEV supplementation. mHFD: neonates born to mHFD dams that did not receive MEV supplementation. mHFD-MEV: neonates born to mHFD dams that received MEV supplementation. n = 3-4 biological replicates/diet/treatment/sex. Data presented are means ± standard error.

HSPA1A transcript abundance in male neonates was responsive to diet (F(_1,12) = 6.179, p = 0.029) and MEV treatment (F(_3,12) = 4.079, p = 0.033) where HSPA1A transcript abundance was higher in mCHD-MEV than mHFD-MEV (Tukey HSD, p = 0.026) (Fig. 7b). HSPA1A transcript in females remained unchanged (diet: (F(_1,12) = 3.343, p = 0.092); treatment (F(_3,12) = 1.642, p = 0.232)) (Fig. 7b).

HSP90AA1 transcript abundance in male neonates was responsive to diet (F(_1,11) = 27.483, p < 0.001) and MEV treatment (F(_3,11) = 13.174, p < 0.001), where HSP90AA1 transcript was higher in mHFD-MEV than mCHD (Tukey HSD, p < 0.001), mCHD-MEV (Tukey HSD, p = 0.002), and mHFD (Tukey HSD, p = 0.028) (Fig. 7c). HSP90AA1 transcript abundance in female neonates was also responsive to diet (F(_1,12) = 25.717, p < 0.001), and MEV treatment (F(_3,12) = 9.613, p = 0.002), where HSP90AA1 transcript was higher in mHFD-MEV than mCHD (Tukey HSD, p = 0.004) and mCHD-MEV (Tukey HSD, p = 0.003) (Fig. 7c).

DNAJB1 transcript abundance in male neonates did not change with diet (F(_1,11) = 4.272, p = 0.063), although a treatment effect was observed (F(_3,11) = 4.364, p = 0.030) where DNAJB1 transcript was higher in mCHD than mHFD (Tukey HSD, p = 0.039) (Fig. 7d). DNAJB1 transcript remained unchanged in female neonates (diet: (F(_1,12) = 0.210, p = 0.655); treatment (F(_3,12) = 1.988, p = 0.170)) (Fig. 7d). Transcript data of male and female neonates that received PBS gavage (the sham controls) is included in Supplementary Table S7 .

Protein abundance of HSR targets in offspring hypothalamus

In the hypothalamus, HSF1 protein abundance in male neonates remained unchanged (diet: (F(_1,10) = 1.627, p = 0.231); treatment (F(_3,10) = 0.985, p = 0.438)) (Fig. 8a). HSF1 protein abundance in female neonates was responsive to diet (F(_1,10) = 10.413, p = 0.009) and MEV treatment (F(_3,10) = 4.180, p = 0.037), where HSF1 protein was higher in mHFD than mCHD (Tukey HSD, p = 0.036) (Fig. 8a).

Figure 8. Protein abundance of the heat shock response targets in the hypothalamus of male and female neonates at postnatal day (PND) 11 as determined by western immunoblotting. (a) HSF1 hypothalamus protein abundance. (b) Hsp70 hypothalamus protein abundance. (c) Hsp90 hypothalamus protein abundance. (d) Hsp40 hypothalamus protein abundance. Protein targets are normalized to the abundance of total soluble proteins in the samples using Coomassie staining. ECL and Coomassie-stained images are displayed. #Main effect of diet (p < 0.05). *Main effect of treatment (p < 0.05). Pairwise comparisons between treatment groups are indicated with lowercase letters, where significant differences (p < 0.05) are denoted by different letters. mCHD: neonates born to mCHD dams that did not receive MEV supplementation. mCHD-MEV: neonates born to mCHD dams that received MEV supplementation. mHFD: neonates born to mHFD dams that did not receive MEV supplementation. mHFD-MEV: neonates born to mHFD dams that received MEV supplementation. n = 3-4 biological replicates/diet/treatment/sex. Data presented are means ± standard error.

Hsp70 protein abundance in male neonates remained unchanged (diet: (F(_1,10) = 3.987, p = 0.074); treatment (F(_3,10) = 2.835, p = 0.092)) (Fig. 8b). Hsp70 protein abundance in female neonates was responsive to diet (F(_1,10) = 20.403, p = 0.001) and MEV treatment (F(_3,10) = 7.324, p = 0.007) where Hsp70 protein was lower in mHFD-MEV than mCHD (Tukey HSD, p = 0.016) and mCHD-MEV (Tukey HSD, p = 0.019) (Fig. 8b).

Hsp90 protein abundance in male neonates remained unchanged (diet: (F(_1,10) = 2.687, p = 0.132); treatment (F(_3,10) = 2.647, p = 0.106)) (Fig. 8c). Hsp90 protein abundance in female neonates also remained unchanged (diet: (F(_1,10) = 1.162, p = 0.306); treatment (F(_3,10) = 1.686, p = 0.232)) (Fig. 8c).

Hsp40 protein abundance in male neonates was responsive to diet (F(_1,10) = 33.566, p < 0.001), and MEV treatment (F(_3,10) = 11.790, p = 0.001), where Hsp40 protein was higher in mCHD compared to mHFD (Tukey HSD, p = 0.008) and mHFD-MEV (Tukey HSD, p = 0.002). Hsp40 protein was also higher in mCHD-MEV compared to mHFD (Tukey HSD, p = 0.043) and mHFD-MEV (Tukey HSD, p = 0.011) (Fig. 8d). Hsp40 protein abundance in female neonates remained unchanged (diet: (F(_1,10) = 4.794, p = 0.053); treatment (F(_3,10) = 2.847, p = 0.091) (Fig. 8d). Protein data of male and female neonates that received PBS gavage (the sham controls) is included in Supplementary Table S7 .

Transcript abundance of HSR targets in offspring prefrontal cortex

In the prefrontal cortex, HSF1 transcript abundance in male neonates remained unchanged (diet: (F(_1,12) = 2.867, p = 0.116); treatment (F(_3,12) = 1.486, p = 0.268)) (Fig. 9a). HSF1 transcript abundance in female neonates also remained unchanged (diet: (F(_1,11) = 0.902, p = 0.363); treatment (F(_3,11) = 0.920, p = 0.463)) (Fig. 9a).

Figure 9. Transcript abundance of the heat shock response genes in the prefrontal cortex of male and female neonates at postnatal (PND) 11 as determined by RT-qPCR. (a) HSF1 prefrontal cortex transcript abundance. (b) HSPA1A prefrontal cortex transcript abundance. (c) HSP90AA1 prefrontal cortex transcript abundance. (d) DNAJB1 prefrontal cortex transcript abundance. Quantity means are normalized to the geometric mean of two reference genes with stable expression: 18S rRNA and YWAZ. #Main effect of diet (p < 0.05). *Main effect of MEV treatment (p < 0.05). Pairwise comparisons between treatment groups are indicated with lowercase letters, where significant differences (p < 0.05) are denoted by different letters. mCHD: neonates born to mCHD dams that did not receive MEV supplementation. mCHD-MEV: neonates born to mCHD dams that received MEV supplementation. mHFD: neonates born to mHFD dams that did not receive MEV supplementation. mHFD-MEV: neonates born to mHFD dams that received MEV supplementation. n = 3-4 biological replicates/diet/treatment/sex. Data presented are means ± standard error.

HSPA1A transcript abundance in male neonates remained unchanged (diet: (F(_1,12) = 0.565, p = 0.467); treatment (F(_3,12) = 1.918, p = 0.181)) (Fig. 9b). HSPA1A transcript abundance also remained unchanged in female neonates (diet: (F(_1,11) = 1.416, p = 0.259); treatment (F(_3,11) = 1.016, p = 0.423)) (Fig. 9b).

HSP90AA1 transcript abundance in male neonates remained unchanged (diet: (F(_1,12) = 0.717, p = 0.414); treatment (F(_3,12) = 0.328, p = 0.805)) (Fig. 9c). HSP90AA1 transcript abundance also remained unchanged in female neonates (diet: (F(_1,12) = 0.010, p = 0.923); treatment (F(_3,12) = 0.044, p = 0.987)) (Fig. 9c).

DNAJB1 transcript abundance in male neonates remained unchanged (diet: (F(_1,12) = 2.620, p = 0.131); treatment (F(_3,12) = 2.537, p = 0.106)) (Fig. 9d). DNAJB1 transcript abundance also remained unchanged in female neonates (diet: (F(_1,12) = 4.292, p = 0.060); treatment (F(_3,12) = 3.002, p = 0.073)) (Fig. 9d). Transcript data of male and female neonates that received PBS gavage (the sham controls) is included in Supplementary Table S7 .

Protein abundance of HSR targets in offspring prefrontal cortex

In the prefrontal cortex, HSF1 protein abundance remained unchanged in male neonates (diet: (F(_1,10) = 0.532, p = 0.483); treatment (F(_3,10) = 2.634, p = 0.107)) (Fig. 10a). HSF1 protein abundance in female neonates did not change with diet (F(_1,10) = 2.414, p = 0.151), but there was an effect of treatment (F(_3,10) = 6.884, p = 0.009), where HSF1 protein was higher in mHFD-MEV compared to mCHD-MEV (Tukey HSD, p = 0.008) and mHFD (Tukey HSD, p = 0.044) (Fig 10a).

Figure 10. Protein abundance of the heat shock response targets in the prefrontal cortex of male and female neonates at postnatal day (PND) 11 as determined by western immunoblotting. (a) HSF1 prefrontal cortex protein abundance. (b) Hsp70 prefrontal cortex protein abundance. (c) Hsp90 prefrontal cortex protein abundance. (d) Hsp40 prefrontal cortex protein abundance. Protein targets are normalized to the abundance of total soluble proteins in the samples using Coomassie staining. ECL and Coomassie-stained images are displayed. #Main effect of diet (p < 0.05). *Main effect of treatment (p < 0.05). Pairwise comparisons between treatment groups are indicated with lowercase letters, where significant differences (p < 0.05) are denoted by different letters. mCHD: neonates born to mCHD dams that did not receive MEV supplementation. mCHD-MEV: neonates born to mCHD dams that received MEV supplementation. mHFD: neonates born to mHFD dams that did not receive MEV supplementation. mHFD-MEV: neonates born to mHFD dams that received MEV supplementation. n = 3–4 biological replicates/diet/treatment/sex. Data presented are means ± standard error.

Hsp70 protein abundance in male neonates did not change with diet (F(_1,10) = 0.383, p = 0.550) but there was an effect of treatment (F(_3,10) = 6.703, p = 0.009), where Hsp70 protein was lower in mHFD-MEV than mCHD (Tukey HSD, p = 0.024) and mHFD (Tukey HSD, p = 0.026) (Fig. 10b). Hsp70 protein abundance remained unchanged in female neonates (diet: (F(_1,10) = 0.096, p = 0.763); treatment (F(_3,10) = 0.208, p = 0.889) (Fig. 10b).

Hsp90 protein abundance in male neonates did not change with diet (U = 20.000, p = 0.945), but there was a treatment effect (H(3) = 8.170, p = 0.043), where MEV supplementation reduced the protein levels of Hsp90 compared to the respective controls. Specifically, mCHD-MEV was lower than mCHD (Dunn’s post hoc, p = 0.046) and mHFD (Dunn’s post hoc, p = 0.046). Similarly, mHFD-MEV was lower than mCHD (Dunn’s post hoc, p = 0.041) and mHFD (Dunn’s post hoc, p = 0.041) (Fig. 10c). Hsp90 protein abundance in female neonates was responsive to diet (F(_1,10) = 16.320, p = 0.002), and MEV treatment (F(_3,10) = 11.946, p = 0.001), where Hsp90 protein was lower in mCHD-MEV than mCHD (Tukey HSD, p = 0.027), mHFD (Tukey HSD, p < 0.001), and mHFD-MEV (Tukey HSD, p = 0.022) (Fig. 10c).

Hsp40 protein abundance in male neonates was responsive to diet (F(_1,10) = 45.741, p < 0.001), and MEV treatment (F(_3,10) = 22.029, p < 0.001) where Hsp40 protein was lower in mCHD-MEV than mHFD-MEV (Tukey HSD, p = 0.012) (Fig. 10d). Additionally, Hsp40 protein was higher in mHFD than mCHD (Tukey HSD, p = 0.001), mCHD-MEV (Tukey HSD, p < 0.001) and mHFD-MEV (Tukey HSD, p = 0.010). Hsp40 protein abundance in female neonates remained unchanged (diet: (F(_1,10) = 0.559, p = 0.472); treatment (F(_3,10) = 1.585, p = 0.254)) (Fig. 10d). Protein data of male and female neonates that received PBS gavage (the sham controls) is included in Supplementary Table S7 .

Discussion

We investigated the interactions between MEVs and the HSR in the liver, hypothalamus, and prefrontal cortex in male and female neonatal rats exposed to mHFD and mCHD at PND11. PND11 is a critical window of early development in Long Evans rats, representing the peak lactational period and within the stress hyporesponsive period (SHRP) that spans the first 14 days of postnatal life. ^72–74 Although acute stress that leads to heightened glucocorticoid release is reduced during the SHRP, ⁷⁵ chronic stress such as mHFD, leads to significant developmental and inflammatory outcomes in offspring. ^5,17,56,57 The HSP chaperones involved in the HSR participate in protein folding and refolding mechanisms, disaggregation, and degradation. ⁷⁶ Our overall findings suggests target, tissue, and sex-specific interactions between MEVs and the HSR, where MEV treatment to offspring with perinatal mHFD is associated with upregulated HSF1 expression and downregulated expression of its inhibitors in the prefrontal cortex. Further, females were likely more responsive to MEV treatment than male littermates.

We observed minimal differences in bodyweight between mCHD and mHFD dams during pre-gestation, gestation, and lactation (Fig. 2a). Previous studies have reported that dams on this particular high-fat diet exhibit higher bodyweights than mCHD dams. ^17,57 However, these studies used a rodent chow as the control diet and did not use the same mCHD, which is matched in sucrose content to the mHFD. Body weight gain is directly proportional to the dietary composition and palatability in rats with diet-induced obesity. ⁷⁷ Nevertheless, as shown by Abuaish et al. ⁵⁷ and Sasaki et al. ¹⁷ the mHFD dams consumed more kcal than mCHD dams during pre-gestation (Fig. 2e).

Our offspring data corroborates previous studies that have used the same mHFD to induce perinatal diet stress in Long Evans neonates, ^17,57 where mHFD offspring were significantly heavier than mCHD offspring postnatally (Fig. 3a). The Lee index, a measure of obesity in rodents that accounts for naso-anal length and bodyweight, ⁶⁷ remained unchanged across diets (Fig. 3b). Correspondingly, at PND11, mHFD offspring had significantly higher amounts of retroperitoneal fat than mCHD offspring, as well as significantly higher amounts of stomach milk curd (Fig. 4a, 4c). In combination, our data shows that mHFD exposure during perinatal life may lead to increased bodyweight and adiposity in both males and female neonates and these changes occur during the SHRP. We did not find sex or MEV treatment effects in bodyweight and Lee Index.

The liver is a multicellular organ that plays significant roles in metabolism. ^78,79 Multiple studies have reported that small extracellular vesicles bioaccumulate in the liver following administration, before being repackaged and re-distributed to other organs. ^59,80–83 Furthermore, the liver is also impacted by pro-inflammation via the liver-gut-brain axis ³¹ with perinatal mHFD exposure. ^84,85 As such, we investigated if and how MEVs may regulate the HSR in neonatal liver in response to chronic diet stress. We found a strong sex effect in HSR at the transcript and protein levels, where HSR in females was more responsive compared to male littermates. These findings are consistent with previous studies that have found females to be more sensitive to perinatal nutritional stress than males during the SHRP. ^5,17,57 At the transcript level, females with perinatal mCHD responded to MEV treatment more robustly than males, with increased HSF1, HSPA1A, and DNAJB1 (Fig. 5a, 5b, 5d). No changes were seen at the protein level across sex, diet, and treatment (Fig. 6), with exception to Hsp40, where mHFD males had lower Hsp40, and mHFD females had higher Hsp40 (Fig. 6d). Further studies are required to determine if this inverse regulation in Hsp40 between sexes influences overall cytoprotection and capability to withstand and respond to proteotoxic stress in early life. In addition, transcript-level changes in HSR in mCHD offspring did not correlate with protein abundance. It is possible that these responses are not being translated to the protein level because the baseline protein levels of the candidate HSPs are sufficient to combat proteotoxic stress and additional increases in cytoprotection are not needed. Indeed, distribution patterns and expression of Hsp70 and Hsp40 protein and RNA, ^86–89 summarized in the Human Protein Atlas (proteinatlas.org), are typically shown to be higher in the liver compared to the CNS.

The hypothalamus is responsible for thermoregulation, circadian rhythm, feeding behavior, energy metabolism, social responses and is a central component of the limbic system. ^24,25 The PVN of the hypothalamus directly regulates HPA axis activity. ⁹⁰ In the hypothalamus, HSP90AA1 transcript decreased in males with MEV treatment (Fig. 7c). The only other effects were diet-induced, where mHFD increased HSF1 and HSP90AA1, and decreased HSPA1A at the transcript level (Fig. 7a–7c), and increased HSF1 and decreased Hsp70 and Hsp40 at the protein level (Fig. 8a, 8b, 8d). Given that the hypothalamus plays a central role in regulating feeding behavior, HPA axis regulation, and is sensitive to diet-induced pro-inflammation, it is unsurprising that we primarily see diet effects. De Souza et al. ⁹¹ analyzed the hypothalamus of male Wistar rats following 16-week consumption of a standard diet (10% kcal lipids) or hyperlipidic diet (45% kcal lipids). They found that a hyperlipidic diet induces the expression of several pro-inflammatory proteins including TNFα, IL-1β, and IL-6. Furthermore, they reported that a hyperlipidic diet results in increased c-Jun N-terminal kinase and NFκB activation, corroborating the sensitivity of the hypothalamus to diet stress. Likely diet resulted in heightened baseline levels in mHFD exposed offspring, and MEVs were unable to exert strong cytoprotective effects. Furthermore, the hypothalamus is a heterogeneous brain region with distinct nuclei which are spatially separated into the preoptic, tuberal and posterior regions. ^92,93 Given its heterogeneity, it is likely that MEV effects are difficult to delineate in the whole hypothalamus, as opposed to distinct nuclei within the hypothalamus, such as the PVN. Huang et al. ⁹² analyzed single nucleus transcriptomics in the hypothalamus of PND15 C57BL/6 mice exposed to mCHD (10% kcal fat) or mHFD (45% kcal fat). They found that within the 30 cellular subpopulations analyzed within the hypothalamus, neuronal subpopulations respond differentially to mHFD, especially for markers involved in feeding behavior, glucose homeostasis, insulin signaling, and circadian rhythm.

The prefrontal cortex is involved in higher-order executive and cognitive function, forming extensive connections with other brain regions, including the hypothalamus and amygdala. ^94,95 It is involved in glucocorticoid receptor-mediated feedback inhibition of the HPA axis. ^23,28–30 At the transcript level in the prefrontal cortex, MEV treatment was not associated with changes in HSR targets, irrespective of diet and sex (Fig. 9a–9d). At the protein level, there were significant differences in mHFD offspring, where protein abundance of Hsp70, Hsp90 and Hsp40 decreased with MEV treatment (Fig. 10b–10d), while HSF1 increased with MEV treatment (Fig. 10a). Previous studies corroborate our findings and report that increased HSF1 activity in the brain, including the prefrontal cortex and hippocampus, is likely neuroprotective against chronic stress. ^96–98 Furthermore, Hsp70 and Hsp90, members of the Hsp70-Hsp90 protein refolding complex, are implicated in the negative regulation of HSF1. ⁹⁹ Hsp70 limits the hyperactivation and prolonged function of HSF1 by interfering with its binding and stabilization to keep it in a structurally inactive form. ⁹⁹ Concomitantly, Hsp90 disassembles structurally active HSF1, attenuating HSF1-mediated gene expression of HSP chaperones. ⁹⁹ Hsp40 acts as a co-chaperone that enhances the ATPase activity of Hsp70 and thus is typically found to follow the same regulatory patterns as Hsp70. ^100,101 Our results illustrate that MEV treatment in the prefrontal cortex is associated with upregulation of HSF1, and downregulation of the negative regulators in male and female offspring with perinatal mHFD exposure. Furthermore, our results may suggest that MEV treatment is prolonging the HSR, by maintaining HSF1 transcription. We reported a similar regulation in vitro using immortalized human microglia clone 3 (HMC3) cell line, where MEV treatment increases protein abundance of HSF1, while decreasing the abundance of the Hsp70 and Hsp90 negative regulators. ⁵⁵ One of the limitations of the study is the small sample size (n = 4 animals/diet/treatment/sex). Although, the current sample size falls within the minimum power requirements as per Arifin et al. ⁶⁸ for group comparisons and because it is vital for ethical reasons to keep the sample size of animals at the lowest sufficient level to adhere to the 3Rs of animal research (Replacement, Reduction, and Refinement), future studies should expand the number of biological samples to reduce the Type II Error. Note: this study was originally designed to include n = 6 animals/diet/treatment/sex across 12 individual litters. Unfortunately, due to reproductive limitations of select dams, we were only able to generate n = 8 litters with matching sex ratios (66.7% success for mCHD and mHFD).

Conclusion

The therapeutic potential of MEVs increases as their role in improving the progression of various pro-inflammatory disease and afflictions is continually discovered. More recently, MEVs have been shown to cross the blood-brain barrier, delivering their cargo to regions within the brain. However, despite this knowledge, the cytoprotective potential of MEVs within the brain remains largely understudied. Our study is the first to investigate interactions between MEVs and the HSR in key tissues and regions involved in HPA axis-mediated stress responses, in the context of promoting cytoprotection in response to mHFD exposure. We found that the HSR was more robust in female neonates than males with MEV treatment. Through our perinatal mHFD model, we found that while MEV treatment was associated with transcript-level changes in mCHD female offspring in the liver, and diet effects in the hypothalamus, stronger effects were evident in the prefrontal cortex. Specifically, MEV supplementation during critical periods of early postnatal life was associated with an increased abundance of the main transcription factor, HSF1, in the prefrontal cortex, while also downregulating its negative regulators, the Hsp70-Hsp90 complex, and Hsp70-binding partner Hsp40. We postulate that by downregulating Hsp70 and Hsp90, HSR activation is potentially prolonged by MEV treatment, resulting in continued pro-survival benefits in response to mHFD stress. Future research will focus on investigating components of the MEV cargo inducing these effects as well as other downstream pro-survival targets regulated by HSF1 that may be activated.