HFD led to impaired fear extinction in male but not female mice

The study’s experimental design is illustrated in Fig. 1A. First, we assessed the percentage of freezing behavior of female and male mice during SEFL paradigm before the dietary regimen. To determine the possibility of fear generalization on Day 2 of SEFL, we assessed percentage of freezing behavior in female and male mice. There were no observed differences between the groups (Fig. 1B) (n = 19,20; two-tailed unpaired t test, t = 1.913, p = 0.0635). We also observed no difference in shock reactivity between the groups to the foot shock delivered in context B on day 2 (Fig. 1C) (n = 19,20; two-tailed unpaired t test, t = 1.875, p = 0.0687). Our study revealed no statistically significant differences between female and male mice in terms of the percentage of freezing on day 3 of SEFL paradigm before the dietary regimen (Fig. 1D) (n = 19,20; two-tailed unpaired t test, t = 1.814, p = 0.0774). However, during the 14-week period, we found a main effect of stress on HFD-fed male mice that were subjected to repeated shocks (Fig. 1E n = 4,5; two-way ANOVA; F1,98 =69.18, p < 0.05). Post hoc comparisons revealed that by week 14, there were no significant differences in freezing behavior between male mice subjected to HFD-NRS or HFD-RS (n = 4,5; Sidak’s multiple comparison test; p > 0.99) (Fig. 1E). These findings indicate that male mice fed a HFD did not exhibit fear extinction by week 14, as evidenced by the consistently high percentage of freezing. This persistent response was notably also not found to be statistically different from HFD-RS male mice during the freezing test conducted after the 14-week period (Fig. 1F) (n = 4,5; two-tailed paired t test; t = 2.518, p = 0.0864). Two-way ANOVA of HFD-fed female mice revealed a main effect of stress (n = 5; two-way ANOVA, F1,112 =157.1, p < 0.05). Post hoc comparisons revealed a significant reduction in freezing percentage in the NRS group compared to that in the RS group at week 14 (Fig. 1G) (n = 5; Sidak’s multiple comparison test; p < 0.05). Similarly, NRS-female mice exhibited a significant reduction in % freezing during the freezing test conducted at week 14 (Fig. 1H) (n = 5; two-tailed paired t test; t = 3.514, p < 0.05).

Fig. 1

A high-fat diet inhibits fear extinction in male mice. (A) Experimental timeline used to study the effect of diet and repeated stress on mice. (B) Percent freezing on day 2 of SEFL (n = 20/19; two-tailed unpaired t test, t = 1.913; p = 0.0635). (C) Shock reactivity measured as velocity (cm/s) on day 2 (n = 20/19; two-tailed unpaired t test, t = 1.875; p = 0.0687). (D) Percent freezing on day 3 of SEFL (n = 20/19; two-tailed unpaired t test, t = 1.814; p = 0.3916). (E) Percent freezing of HFD-fed male mice during 14 weeks of RS/NRS treatment (n = 4,5; two-way ANOVA, F1,98 =69.18; ****p < 0.0001; post hoc comparison, n = 5/4; Sidak’s multiple comparisons test, p > 0.99). (F) The freezing percentage of HFD-fed male mice during the freezing test conducted after 14 weeks (n = 5/4, two-tailed paired t test; t = 2.518, p = 0.08). (G) Percent freezing of HFD female mice during 14 weeks of RS/NRS treatment (n = 5; two-way ANOVA, F1,112 =157.1; ****p < 0.0001) (post hoc comparison, n = 5; Sidak’s multiple comparisons test, **p < 0.01). (H) Percent freezing in HFD females in the freezing test after 14 weeks (n = 5, two-tailed paired t test; t = 3.514, *p < 0.05)

In the chow fed male and female mice groups, we observed a reduction in % freezing by week 14 (n = 5 males; two-way ANOVA; F1,112 =89.76, p < 0.05) (n = 5 males, Sidak’s multiple comparison test, p < 0.0001) (n = 5 females, two-way ANOVA; F1,112 =161.1, p < 0.05) (n = 5 females, Sidak’s multiple comparison test; p < 0.0001). (Sup. Figure S1 A & S1 C) in the chow-NRS group compared to that in the chow-RS groups. Additionally, we observed a significant decrease in the percentage of freezing behavior during the freezing test conducted at week 14 in both male and female mice (Sup. Figure S1B & S1D) (n = 5 males, two-tailed paired t test; t = 3.586, p = 0.023). (n = 5 females, two-tailed paired t test; t = 4.771, p = 0.008). In the open field light gradient tests conducted after, there were no main effects of diet or sex in any of the HFD/chow groups (n = 4,5, two-way ANOVA; p > 0.05) (Supplemental Fig. 1E-1G).

HFD and repeated stress induced weight gain and disrupted blood glucose homeostasis in both male and female mice

We observed that male and female mice were significantly different in percentage of weight gain under HFD (n = 4,5; two-way ANOVA; F 32,235 =37.12, p < 0.0001) (Fig. 2A). Post hoc comparisons revealed a statistically significant difference in weight gain between male and female mice within the HFD-fed groups (n = 4,5; Tukey’s multiple comparison test; p < 0.0001). Similarly, as seen in Fig. 2B, male and female mice fed a Chow diet displayed a significant differences in % weight gain with males increasing weight at a higher rate than females (n = 5; two-way ANOVA; F 3,240 =5.698, p = 0.009). Post hoc comparison revealed a significant difference between chow-RS-F and chow-NRS-F and between chow-RS-F and chow-RS-M (n = 5, Tukey’s multiple comparison test, p = 0.0097). Next, we examined whether repeated stress and a HFD influenced glucose tolerance and insulin sensitivity through a glucose tolerance test (GTT) and an insulin tolerance test (ITT), respectively. Our results revealed glucose levels were significantly different in male mice and female mice fed HFD (Fig. 2C) (n = 5; two-way ANOVA; F 3,90 =33.23, p < 0.0001). Post hoc analysis revealed that the differences were significant at the 60- to 90-minute time points in the HFD-female group compared with the HFD-male group (n = 5; Tukey’s multiple comparison test; p < 0.05, p < 0.01, p < 0.001, p < 0.0001). Furthermore, the area under the curve (AUC) plot also revealed differences between male and female mice (Fig. 2D) (n = 5; one-way ANOVA; F 3,15 =9.531, p = 0.0009). Post hoc comparison revealed differences between male and females fed HFD irrespective of stress grouping (n = 4,5, Tukey’s multiple comparison test, p < 0.05, p < 0.01). By week 14, we also observed a significant effect of sex and stress on blood glucose levels (n = 5, two-way ANOVA; F 3,90 =14.91, p < 0.0001). Post hoc analysis confirmed a significantly elevated blood glucose level in male mice in the HFD-RS and HFD-NRS group, particularly at the 30- to 90-minute time points, compared to that in the HFD-NRS female groups (n = 5; Tukey’s multiple comparison tests; p < 0.05). Subsequently, female mice in the HFD-RS group exhibited heightened blood glucose levels and no longer showed blunted glucose levels in comparison to HFD-male mice. Further, we observe a sex effect in the AUC graph (n = 5; one-way ANOVA; F 3,15 =4.605, p = 0.017) (Fig. 2D). Post hoc comparison shows significant differences between HFD-NRS-F and both HFD male groups (n = 4,5, Tukey’s multiple comparison test, p = 0.03). Subsequently, we assessed blood glucose levels at 10 and 14 weeks in male and female mice from the chow groups in the NRS and RS. At week 10, we observed a distinct effect of sex and stress on blood glucose levels in chow fed mice (Fig. 2E) (n = 4,5; two-way ANOVA; F 3,90 =49.41, p < 0.0001). Post hoc analysis indicated that the differences started to increase at the 30-minute time point and were sustained until the 120th minute of measurement (n = 4,5; Tukey’s multiple comparison tests; p < 0.05, p < 0.01, p < 0.001, p < 0.0001). The area under the curve (AUC) further showed a significant sex and stress effect (n = 4,5; one-way ANOVA; F 3,16 =4.288, p = 0.0212). At week 14, however, we found no significant effect of the sex or stress on the GTT in male or female mice fed chow (Fig. 2F) (n = 4,5; two-way ANOVA; F 3,96 =59.52, p = 0.618). The area under the curve (AUC) graph further illustrates no difference between the groups (n = 4,5; one-way ANOVA; F 3,16 =1.867, p = 0.1758).

Fig. 2

HFD induces weight gain and an increase in blood glucose levels in a sex-specific manner. (A) Percent weight gain over 14 weeks in HFD-fed mice (n = 4/5, two-way ANOVA; F 3,225 =37.12; ****p < 0.0001) (Post hoc comparison, n = 4/5, Tukey’s multiple comparisons test, p < 0.05 from week 7). (B) Percent weight gain over 14 weeks plotted for chow-fed mice (n = 5; two-way ANOVA, F 3,240 =5.698, ***p < 0.001) (Post hoc comparison, n = 5, Tukey’s multiple comparisons test, p < 0.05 in week 10 only). (C) Week 10 plasma glucose levels during the GTT test after 4 h of fasting in HFD mice (n = 5, two-way ANOVA, F 3,90=33.23, ****p < 0.0001) (post hoc comparison, n = 5, Tukey’s multiple comparisons test, p < 0.05). AUC graph from the 10-week GTT test in HFD mice (N = 5, one-way ANOVA, F 3,15 =1.579, ***p < 0.001) (post hoc comparison, N = 5, Tukey’s multiple comparisons test, p < 0.05) (D) Week 14 plasma glucose levels during the GTT test in HFD mice (n = 5, two-way ANOVA, F 3,90=14.91, ****p < 0.0001) (post hoc comparison, N = 5, Tukey’s multiple comparisons test, p < 0.05). AUC graph from the 14-week GTT test in HFD mice (n = 5, one-way ANOVA; F 3,15=0.8293, *p < 0.05) (post hoc comparison, n = 5; Tukey’s multiple comparison test; p < 0.05). (E) Week 10 plasma glucose levels during the GTT test in chow-fed mice (n = 5, two-way ANOVA, F 3,96=12.54, ****p < 0.0001) (post hoc comparison, n = 4/5, Tukey’s multiple comparisons tests, p < 0.05). AUC graph from the 10-week GTT test in chow-fed mice (n = 5, one-way ANOVA; F 3,16=0.0884, *p < 0.05) (post hoc comparison, n = 4/5, Tukey’s multiple comparisons test, p < 0.05). (F) Week 14 plasma glucose levels during the GTT test in chow-fed mice (n = 5, two-way ANOVA, F 3,96=6.201, ***p < 0.001) (post hoc comparison, n = 4/5, Tukey’s multiple comparisons tests, p = 0.08). AUC graph from the 14-week GTT test in chow-fed mice (n = 5, one-way ANOVA; F 3,65=0.075; p = 0.59) (*Compares No repeated Shock F vs. No repeated Shock M, #Compares No repeated shock F to Repeated Shock M, +Compares Repeated shock F to No repeated Shock M, %Compares Repeated shock F to Repeated Shock M)

We conducted ITTs on 14-week-old HFD-fed, chow-fed female and male mice subjected to either NRS or RS. We did not see any effect of stress or sex in HFD fed mice (Sup. Figure 2 A) (n = 5; two-way ANOVA; F 3,16 =1.438, p = 0.38) further confirmed by the area under the curve (AUC) (n = 5; one-way ANOVA; p = 0.595). However, in chow-fed mice, we see a significant effect of sex (Sup. Figure 2B) (n = 4,5; two-way ANOVA; F 3,96 =112.1, p < 0.0001). Post hoc analysis revealed that, compared with chow-fed female mice, chow-fed male mice exhibited higher blood glucose revealing higher insulin sensitivity at all time points (n = 4,5; Tukey’s multiple comparison tests, p < 0.05, p < 0.01, p < 0.001, p < 0.0001). The area under the curve (AUC) further confirmed the observed differences in ITTs between the diet groups of male mice (n = 5; one-way ANOVA; F 3,16 =68.94, p < 0.0001).

Differential effects of HFD and repeated stress on energy balance

To explore the impact of diet and repeated shocks on energy metabolism, we subjected both chow-fed and HFD-fed mice from both the RS and NRS groups to a 72-hour period in metabolic chambers. This approach allowed us to measure indirect calorimetry parameters, including energy expenditure (EE), the respiratory exchange ratio (RER), food intake, and locomotion. In our analysis, we excluded the initial 12 h to allow the mice to acclimatize. We observed a main effect of sex on the energy expenditure of mice fed a HFD (Fig. 3A) (n = 4,5; one-way ANOVA; p < 0.05). Post hoc comparison revealed that under HFD conditions, male mice exhibited significantly greater EE than female mice. However, EE was comparable between the RS and NRS groups within both male and female mice, indicating that there was no discernible stress effect on EE (n = 4,5; Tukey’s multiple comparison test; p < 0.05). In the context of a chow diet, there was no discernible impact of sex or stress on overall EE (Fig. 3B) (n = 5; one-way ANOVA; p > 0.05). Subsequently, we investigated fuel utilization patterns in both male and female mice across various diets and stress conditions by assessing RER levels. The RER, denoting the ratio of carbon dioxide (CO2) production to oxygen (O2) absorption, serves as an indicator of the specific energy source—carbohydrate or fat—that the body metabolizes to fulfill its energy requirements. We observed a notable main effect of sex on the RER in HFD-fed mice (Fig. 3C) (n = 4,5; one-way ANOVA; p < 0.05). Post hoc comparisons revealed marked differences in RERs between male mice and female mice. Specifically, female mice exhibited significantly lower RERs under both the RS and NRS conditions (n = 4,5; Tukey’s multiple comparison test; p < 0.05). Our observations indicated the absence of any discernible sex or stress effects on RER values in the animals fed a chow diet (Fig. 3D) (n = 5; one-way ANOVA; p > 0.05). Interestingly, our observations revealed a significant main effect of sex on locomotor activity in mice fed a HFD (Fig. 3E) (n = 5; one-way ANOVA; p < 0.05). Post hoc comparisons revealed that HFD-RS male mice had significantly lower pedestrian locomotion than the HFD-fed female groups (n = 4,5; Tukey’s multiple comparison test; p < 0.05). We also observed a main effect of sex on pedestrian locomotion among mice fed a normal chow diet (Fig. 3H) (n = 5; one-way ANOVA, p < 0.05). Post hoc analysis revealed that both the RS and NRS groups of female chow-fed mice exhibited greater locomotor activity than the male groups (n = 4,5; Tukey’s multiple comparison test, p < 0.05). We did not see any effect of stress or sex in either HFD-fed or chow-fed groups in total food consumption (Sup. Figure 2 C and 2D).

Fig. 3

A HFD and repeated reminder shocks led to sex-specific metabolic dysregulation. (A) Energy expenditure (kCal/hr) of HFD-fed mice over 60 h in the metabolic chamber (n = 4,5; one-way ANOVA, p < 0.05) (post hoc comparison, n = 4,5; Tukey’s multiple comparisons test, *p < 0.05). (B) Energy expenditure (kCal/hr) of chow-fed mice over 60 h in the metabolic chamber (n = 5; one-way ANOVA, p > 0.05). (C) Respiratory exchange ratio (RER) of HFD-fed mice aged more than 60 h in the metabolic chamber (n = 4/5; one-way ANOVA, p < 0.05) (post hoc comparisons, n = 4,5; Tukey’s multiple comparisons test, *p < 0.05). (D) RERs of chow-fed mice aged more than 60 h in the metabolic chamber (n = 5; one-way ANOVA, p > 0.05). (E) Pedestrian locomotion in HFD-fed mice over 60 h in the metabolic chamber (n = 5; one-way ANOVA, p < 0.05) (post hoc comparisons, n = 4,5; Tukey’s multiple comparison test, *p < 0.05). (F) Pedestrian locomotion in chow-fed mice over 60 h in the metabolic chamber (n = 5; one-way ANOVA, p < 0.05) (post hoc comparison, n = 4,5, Tukey’s multiple comparisons test, *p < 0.05)

HFD and acute stress have synergistic impacts on behavioral responses, glucose homeostasis, and energy metabolism

To investigate the impact of HFD on fear-related behaviors and energy metabolism in response to acute stress, we conducted a comprehensive experimental study following a specific timeline and setup (Fig. 4A). As shown in Fig. 4B, our results showed a significant main effect of sex and diet on the weight gain percentage from baseline (n = 10, two-way ANOVA; F3,528 =188.3 & F10,528 =54.75, p < 0.0001). As anticipated, mice fed a HFD exhibited a greater rate of weight gain than control mice fed a chow diet, with discernible differences manifesting as early as the second week of the diet regimen (n = 10; Tukey’s multiple comparison test; p < 0.0001). However, it is noteworthy that male mice in both diet groups exhibited a notably greater weight gain than their female counterparts within the respective diet groups (n = 10, Tukey’s multiple comparison test; p < 0.0001). Next, we examined fat mass in these animals at the end of week 10, and two-way ANOVA showed a main effect of diet (Fig. 4C) (n = 10; two-way ANOVA; F1,48 = 17.69, p < 0.05). Post hoc analysis further revealed that, exclusively among male mice on HFD, there was a significant increase in the percentage of fat mass compared to that of their counterparts on the Chow diet (n = 10, Tukey’s multiple comparison test; p < 0.05).

Fig. 4

A HFD did not affect acute stress-induced fear behaviors but did cause sex-specific metabolic alterations. (A) Experimental design of the comprehensive 10-week study conducted to investigate the role of diet and acute stress. (B) Weight gain percentages plotted through 10 weeks of the HFD/chow diet regimen (n = 10, two-way ANOVA, F3,528 = 188.3, ****p < 0.0001) (Post hoc comparison, n = 10, Tukey’s multiple comparisons test, p < 0.05 from week 2) (C) Percent fat mass in HFD/chow-fed mice (n = 10; two-way ANOVA, F1,48 = 17.69; ***p < 0.001) (post hoc comparison, n = 10, Tukey’s multiple comparisons test, **p < 0.01). (D) Percent freezing in HFD/chow-fed stressed and no stressed groups (n = 5, one-way ANOVA, F7,44 = 1.388, ****p < 0.0001) (post hoc comparison, n = 5, Tukey’s multiple comparison test, Chow-NS-F v. Chow-S-F, Chow-NS-M v. Chow-S-M, HFD-NS-F v. HFD-S-F, HFD-NS-M v. HFD-S-M; ****p < 0.0001). (E) Plasma glucose levels from the GTT performed at week 10 in HFD-NS/S female and male mice (n = 5, two-way ANOVA, F3,168 = 18.53, ****p < 0.0001). AUC graph for the week 10 GTT test in HFD/chow-fed female mice (n = 5, one-way ANOVA; F3,16 = 0.1454, p = 0.07). (F) Plasma glucose levels from the GTT performed at week 10 in Chow–NS/S female and male mice (n = 5, two-way ANOVA, F3,120 = 9.057, ****p < 0.0001). AUC graph for the week 10 GTT test in chow-fed mice (n = 5, one-way ANOVA, F3,16 = 1.206, p = 0.41). (G) Energy expenditure (EE) (kCal/hr) in HFD-fed S and NS mice (n = 5; one-way ANOVA, p > 0.05). (H) Energy expenditure (EE) (kCal/hr) in chow-fed mice under either the S or NS conditions (n = 5; one-way ANOVA, p > 0.05)

The analysis of freezing behavior in these animals revealed a main effect of stressor (Fig. 4D) (n = 5; one-way ANOVA; F7,44 = 1.388, p < 0.05). Post hoc analysis revealed the following differences: Chow-NS-F vs. Chow-S-F, Chow-NS-M vs. Chow-S-M, HFD-NS-F vs. HFD-S-F, HFD-NS-M vs. HFD-S-M (n = 5; Tukey’s multiple comparison test; p < 0.05). However, we do not see any effect of sex or diet.

GTT tests conducted after the acute stress paradigm revealed a significant main effect of sex and stress in HFD-fed mice (Fig. 4E) (n = 5; two-way ANOVA, F3,168 =18.53, p < 0.05). Post hoc comparison further revealed that this sex effect was particularly pronounced during the 60th -120th minute of measurement, specifically between female mice on HFD and male mice on HFD (n = 5, Tukey’s multiple comparison test, p < 0.05). The area under the curve (AUC) did not illustrate significant difference in female and male mice on HFD, however a trend was observable (n = 5; one-way ANOVA, F3,16 =3.084, p = 0.057). Furthermore, in chow mice, we observed a main effect of sex and stress on blood glucose levels (Fig. 4F) (n = 5; two-way ANOVA, F3,96 =12.61, p < 0.05). Post hoc comparison revealed an increase in blood glucose levels in male mice compared to female mice, particularly evident during the 60th -minute interval of measurement (n = 5, Tukey’s multiple comparison test, p < 0.05). The AUC for the GTT did not show any discernible effect of sex or stress on blood glucose levels, however, we do observe a trend (n = 5; one-way ANOVA, F3,16 =1.026, p = 0.059). Subsequently, we placed the animals in metabolic chambers to quantify changes in metabolic parameters. We observed an effect of sex and stress in EE between male and female mice fed HFD S or NS conditions (Fig. 4G) (n = 5; one-way ANOVA, p < 0.05). Post hoc comparison showed a difference between HFD-S-M and HFD-NS-F groups during dark and light phase. We did not observe a sex effect on EE in mice fed a chow diet (Fig. 4H) (n = 5; one-way ANOVA, p > 0.05). We further examined additional metabolic parameters, including the RER, total food consumption and locomotor activity, we observed no main effect of sex or stress on the RER or food intake (Sup. Figure 3 A-3D) (n = 5; one-way ANOVA, p > 0.05). However, we did observe an effect of sex and stress in both HFD and chow groups in locomotor activity (Sup. Figure 3E-3 F) (n = 5; one-way ANOVA, p < 0.05). Post hoc comparison showed group differences between male mice in both stress conditions in comparison to female mice in NS groups.

Repeated stress impacts myeloid lineage cells within the bone marrow and inflammatory cells in the blood, gWAT, aorta, and heart

To investigate the potential for systemic inflammation resulting from repeated shocks, we initially assessed the expression of proinflammatory genes in the gWAT and liver via quantitative polymerase chain reaction (qPCR). We did not observe any discernible effects of stress or sex on C-C Motif Chemokine Ligand 2 (CCL2) gene expression (n = 9,10; two-way ANOVA, p = 0.15) (Sup. Figure 4 A & 4E). Interestingly, we observed a main effect of sex and stress on the gene expression levels of the proinflammatory cytokine interleukin-1beta (Il-1β) (n = 9,10; two-way ANOVA, F1,15 = 9.796; p = 0.0069) and Il-12p40 (n = 9,10; two-way ANOVA, F1,15 = 8.352; p = 0.011) in HFD-fed mice (Sup. Figure 4B & 4 C). According to our analysis of the liver, sex and stress had no discernible effects on the gene expression levels of CCL2, Il-1β, Il-12p40 or tumor necrosis factor-alpha (TNF-α) (n = 9; two-way ANOVA, p > 0.05) (Sup. Figure 4E-4 H).

Inflammation is an important contributor to metabolic abnormalities and collectively plays a role in the progression of cardiometabolic disorders [68]. Although we did not assess the cardiometabolic disorders phenotype in the present study, the examination of leukocyte abundance in organs pertinent to cardiometabolic disorders provides valuable insights into the potential relevance of our findings to this disease. To address this aspect, we utilized flow cytometric analysis (FACS) to analyze cells collected from the bone marrow, blood, gWAT, aorta and heart tissues of HFD-fed male and female mice that received either RS or NRS, as outlined in the methods section. The gating strategies used for flow cytometry data analysis are shown in Sup. Figure 5A-5E. Multipotent progenitors (MPPs) within the bone marrow, including MPP2 and MPP3, serve as precursors that give rise to both myeloid and lymphoid cell populations. We found a main effect of sex on the MPP2 cell population in the bone marrow of HFD-fed mice (n = 9; two-way ANOVA, F1,14 = 290.1, p < 0.0001) (Fig. 5A). Post hoc comparisons revealed a significant reduction in the MPP2 cell population in HFD-RS female mice compared to HFD-NRS male mice (n = 5; Tukey’s multiple comparison test, p = 0.022) (Fig. 5A). Additionally, we observed a significant decrease in MPP2 cell expression in HFD-fed male mice compared to HFD-fed female mice, irrespective of the presence of stress (n = 4/5; Tukey’s multiple comparison test, p < 0.0001). We observed a comparable main effect of sex on the population of MPP3s within the bone marrow of HFD-fed mice (n = 9; two-way ANOVA, F1,14 = 34.48, p < 0.0001) (Fig. 5A). Although a post hoc analysis revealed a decreasing trend in HFD-RS female mice compared to HFD-NRS female mice, the difference was not statistically significant (n = 5, Tukey’s multiple comparison test, p = 0.08). There was a significant reduction in MPP3 cell expression in male mice fed a HFD compared to female mice fed the same diet, irrespective of the stress group (n = 4/5, Tukey’s multiple comparison test, p = 0.003).

Fig. 5

Repeated reminder shocks and a high-fat diet (HFD) induced distinctive alterations in peripheral myeloid lineage cells and inflammatory markers, exhibiting a sexually dimorphic pattern. (A) Quantification of the bone marrow population of MPP2 progenitors in HFD-fed mice (n = 9; two-way ANOVA, F1,14 = 290.1; ****p < 0.0001) (post hoc comparison, n = 5; Tukey’s multiple comparison test, *p < 0.05 and ****p < 0.0001). (B) Quantification of the MPP3 progenitor population in the bone marrow of HFD-fed mice (n = 9; two-way ANOVA; F1,14 = 34.48; ****p < 0.0001) (post hoc comparison, n = 5; Tukey’s multiple comparison test, p = 0.08 and ***p < 0.001). (C) Quantification of granulocyte monocyte progenitor (GMP) cells in HFD-fed mice (n = 9; two-way ANOVA, F1,14 = 42.52; ****p < 0.0001) (post hoc comparison, N = 4/5; Tukey’s multiple comparison test, **p < 0.01). (D) Quantification of monocyte-dendritic cell progenitors (MDPs) in the bone marrow of HFD-fed mice (n = 9; two-way ANOVA, F1,14 = 6.017, *p < 0.05) (post hoc comparison, n = 5; Tukey’s multiple comparison test, *p < 0.05). (E) Quantification of neutrophils in the blood of HFD-fed mice (n = 9; two-way ANOVA, p > 0.05). (F) Quantification of Ly6Chi monocytes in the blood of HFD-fed mice (n = 9; two-way ANOVA, p > 0.05). (G) Quantification of neutrophils in the gonadal white adipose tissue (gWAT) of HFD-fed mice (n = 9; two-way ANOVA, F1,14 = 24.65, ***p < 0.001) (post hoc comparison, n = 5; Tukey’s multiple comparison test, *p < 0.05 and **p < 0.01). (H) Quantification of macrophages in the gWAT of HFD-fed mice (n = 9; two-way ANOVA; F1,14 = 6.296; *p < 0.05) (post hoc comparison; n = 4/5; Tukey’s multiple comparison test; *p < 0.05). (I) Quantification of aortic neutrophils from HFD-fed mice (n = 9; two-way ANOVA; F1,14 = 59.59; ****p < 0.0001) (post hoc comparison; n = 4/5; Tukey’s multiple comparison test; ***p < 0.001). (J) Quantification of macrophages in the aortas of HFD-fed mice (n = 4/5; Tukey’s multiple comparison test, p > 0.05). K. Quantification of neutrophils in the hearts of HFD-fed mice (n = 9; two-way ANOVA, p > 0.05). L. Quantification of macrophages in the hearts of HFD-fed mice (n = 9; two-way ANOVA; F1,14 = 32.04; ****p < 0.0001) (post hoc comparison; n = 4/5; Tukey’s multiple comparison test; *p < 0.05 and **p < 0.01)

Next, upon examination of the downstream products of the MPPs, namely, granulocyte-monocyte precursors (GMPs) and monocyte-dendritic progenitor cells (MDPs), in the bone marrow, we observed a statistically significant main effect of sex (Fig. 5C and D). Hematopoietic stem cells, including MPP2s and MPP3s, progress through stages to generate multipotent common myeloid progenitor cells (CMPs). CMPs further differentiate into GMPs and MDPs, ultimately producing mature leukocytes [69]. Quantification of GMPs in HFD-fed mice revealed a main effect of sex (Fig. 5C) (n = 9; two-way ANOVA, F1,14 = 42.52, p < 0.0001). Post hoc comparison revealed a significant decrease in GMPs in HFD-fed female mice compared to their male counterparts (n = 5 − 4; Tukey’s multiple comparison test, p = 0.001); however, we did not observe any stress effects. Conversely, we observed a main effect of stress on the MDP population in the HFD-fed groups (Fig. 5D) (n = 9; two-way ANOVA, F1,14 = 6.017, p = 0.057). Post hoc comparisons revealed a significant reduction in MDP levels in HFD-RS male mice compared to HFD-NRS female mice (n = 5 − 4; Tukey’s multiple comparison test, p = 0.0201).

We did not observe any discernible effects of sex or stress on neutrophils or the Ly6Chi monocyte population in the analyzed blood samples (Fig. 5E and F) (n = 5 − 4; one-way ANOVA, p = 0.72 & p = 0.42). Next, we analyzed neutrophil and macrophage populations, which are crucial components of systemic metabolic regulation, in gWAT, the aorta and heart tissue (Fig. 5G and L). We observed a significant main effect of sex and stress on gWAT neutrophil levels (Fig. 5G) (n = 9; one-way ANOVA, F1,14=24.65; p = 0.0002) and a main effect of sex and stress on the macrophage population (Fig. 5H) (n = 9; one-way ANOVA, F1,14=6.296; p = 0.04). Post hoc comparisons revealed an increase in neutrophil counts in HFD-fed male mice compared to HFD-RS female mice in the RS and NRS groups (n = 5 − 4; Tukey’s multiple comparison test, p = 0.01 & p = 0.0022) and an increase in macrophage counts in HFD-NRS female mice compared to HFD-RS female mice (n = 5; Tukey’s multiple comparison test, p = 0.043). In the aorta, we observed a notable main effect of sex and stress on neutrophil counts, indicating the opposite pattern (n = 9; two-way ANOVA, F1,14 = 59.59, p < 0.0001) (Fig. 5I). Post hoc analysis revealed significantly greater numbers of neutrophils in HFD-fed female mice than in male mice, irrespective of the stress group (n = 5; Tukey’s multiple comparison test, p = 0.0009). We did not observe a main effect of sex or stress on macrophage numbers in the aorta (Fig. 5J) (n = 5 − 4; one-way ANOVA, F114 = 0.00089, p = 0.97). We also did not observe any effect of sex or stress on heart neutrophil counts in HFD-fed mice (Fig. 5K) (n = 9; two-way ANOVA, F1,14=0.491, p = 0.49). However, ANOVA revealed a main effect of sex and stress on the number of macrophages in the heart (Fig. 5L) (n = 9; two-way ANOVA, F1,14=32.04, p < 0.0001). Post hoc analysis revealed that female mice in the HFD-NRS group had significantly lower levels of macrophages than male mice in the HFD-fed male group did (n = 5 − 4; Tukey’s multiple comparison test, p = 0.019 & p = 0.001). Additionally, female mice in the HFD-RS female group had significantly lower levels of macrophages than male mice in the HFD-fed group (n = 5 − 4; Tukey’s multiple comparison test, p = 0.007). We extended our analysis to tissues from chow-fed mice, and our observations indicated that sex or stress had minimal impacts on the myeloid cell lineage population and proinflammatory marker levels (Sup. Figure 5 F-5 M).

Single-nuclei RNA sequencing of the ventromedial hypothalamus revealed sexually dimorphic cellular functions following HFD feeding and repeated stressors

As mentioned before, we directed our efforts toward the analysis of VMH cellular functions under conditions of stress and HFD consumption using single-nuclei RNA sequencing (snRNAseq). We performed snRNA-seq to examine the impact of repeated shocks and a HFD on the VMH of both male and female mice (Fig. 6A). Nuclei were isolated from the entire VMH of male and female mice from the HFD-RS group (pooled from 5 mice each). The single-nucleus suspension (n ~ 10,000) was subjected to snRNA-Seq using the 10X Genomics platform, and the libraries were sequenced with 1 billion dedicated reads per sample. We utilized the 10X Genomics data processing platform and SeuratV3 to generate cell clusters and identities (refer to Materials and methods) (Fig. 6A). To classify VMH populations based on gene expression, we conducted cluster analysis, as illustrated by uniform manifold approximation and projection (UMAP) plots (Fig. 6B). These plots enabled us to distinguish distinct clusters of GABAergic neurons, mature neurons, dopaminergic neurons, oligodendrocyte precursors, astrocyte populations, glutamatergic neurons, oligodendrocyte populations, endothelial cells, and microglia. Notably, based on the marker genes, we identified two populations, astrocytes, and oligodendrocytes (Fig. 6B). Further examination through UMAP and gene network plots revealed that each cluster uniquely expressed marker genes, demonstrating preferential expression in individual clusters (Sup. Figure S5 A & B). To gain insight into sex differences in VMH remodeling, we generated a cumulative UMAP plot of female and male mice (Fig. 6C). Overall, the UMAP indicated comparable qualitative changes in the relative proportions of VMH clusters between male and female nuclei (Fig. 6C). To characterize the differences in cell fraction between male and female mice, we calculated cluster percentages in relation to the overall combined dataset. We observed variations in astrocytes, oligodendrocytes, microglia, GABAergic neurons, and mature neurons (Fig. 6D). Transcriptional analyses based on gene signatures revealed an active transcriptional network involving transcription factors such as NFIB, MEIS1, SOX10, and FOXO1 in glial cells. These findings suggested that a HFD and repeated stress had significant impacts on the transcriptional program of glial cells (Fig. 6E). Subsequent analysis included subsetting of glial cells, including microglia, astrocytes, and oligodendrocytes. This analysis revealed increased percentages of microglia and oligodendrocytes in female VMH samples, marked by Ctss and Mags, and an increased astrocyte population in male VMH samples. Interestingly, the neuropeptide Pomc and the astrocyte marker Gfap were notably more abundant in male astrocytes than in female astrocytes (Fig. 6F). To delineate transcriptional differences in glial cell populations between male and female mice, we performed differential gene expression analysis on oligodendrocytes, microglia, and astrocytes. The volcano plot in Fig. 6G illustrates the highly upregulated genes, including Xist, Hcrt, and Ctnap5a, in the microglia of female VMHs compared to those in male VMHs. Among the astrocytic population, the highly significant genes in the male VMH were Nrg3, Lingo2, and Grid2 (Fig. 6H). In the pooled female VMH cohort, the oligodendrocyte population consisted of genes such as Pmch, Tsix, and Rplp1 (Fig. 6I). To correlate the gene expression profiles with pathway analysis, we performed Gene Ontology (GO) analysis of the DEGs in the overall neuronal population, astrocytes, and microglia; the results were correlated with cell type activation pathways. Overall neuroinflammation was comparable between male and female VMH samples (Fig. 6J). However, male pooled VMH samples presented an increase in App and Ldlr expression, while female VMH samples presented an increase in Lrp1 and Mapt expression, indicative of differential activation of astrocyte lipid metabolism (Fig. 6K). Consistent with previous results, female VMH samples showed overall activation of microglia, with increased expression of Nr1d1, Sty11, and Clu (Fig. 6L). In line with these observations, cell cluster expression associated with genome-wide association study (GWAS) traits was examined using the MAGMA program, which revealed that female mice were more strongly correlated with Alzheimer’s disease frequency than male mice under conditions of repeated stress and a HFD (Sup. Figure 5 C).

Fig. 6

Repeated reminder shock and a high-fat diet (HFD) induced differential changes in the ventromedial hypothalamus (VMH) of male and female mice. (A) Illustration outlining the snRNA-seq procedure utilizing the 10X Genomics platform workflow, which involves isolating nuclei from the VMH of male and female mice fed a HFD (n = 5 pooled). (B) UMAP plot displaying single cells from this study, color coded by cell type, with cell types identified based on the expression of canonical marker genes. (C) UMAP plot of single cells from the VMH cohort, color coded by cell type and segregated by sample. (D) Heatmap depicting the relative fractions of each cell type in each sample. (E) Heatmap illustrating the regulon activity of the indicated transcription factors, indicating the intensity of gene regulation in specific cell types from the VMH of males and females. (F) UMAP plot of the indicated cell types from the VMH cohort separated by sample; adjacent violin plots displaying the expression levels of specific genes are shown. G, H & I. Volcano plot illustrating differentially expressed genes (DEGs) with fold changes plotted against p values; female vs. male in microglia (G), male vs. female in astrocytes (H), and female vs. male in oligodendrocytes (I). Violin plots highlighting the highly upregulated genes. J, K & L. Pathway analysis of DEGs showing neuroinflammation (J), astrocyte activation (K), and microglial activation (L). M. Interactome representation based on DEGs from male and female mice indicating the strength of interaction between specific cell types. N. Heatmap exhibiting the indicated ligand-cell interactions in male and female mice

To further study how neuronal and non-neuronal cells interact in male and female mice under repeated stress and HFD conditions, we performed cell‒cell communication (ligand‒receptor interaction) analysis using CellChat [70]. We observed markedly more active communication between cell types in female VMH samples than in male VMH samples (Fig.