Sex differences in BNST signaling and BNST CRF in fear processing

Abstract

Fear responses to perceived danger are critical for survival, as they prompt the individual to respond to threats and avoid harm. However, excessive fear can impede normal biological processes and become harmful. This study investigates the neural mechanisms underlying two distinct forms of fear-phasic and sustained-in male and female mice, with a focus on the bed nucleus of the stria terminalis (BNST) and corticotropin-releasing factor (CRF) signaling. Phasic fear is characterized by immediate responses to clear threats, while sustained fear is driven by ambiguous or uncertain cues and persists longer. Using rodent models, we found that sustained fear, modeled by partial fear conditioning, induced greater arousal and BNST activity in males, especially during ambiguous threat cues. In contrast, females exhibited reduced BNST and BNST^CRF activity, highlighting significant sex differences in fear learning and expression. Additionally, CRF is crucial for appropriate fear response in females, as CRF knockdown led to increased fear responses, but had no effect in males. These sex-specific differences could help inform the development of targeted treatments for anxiety and trauma-related disorders, which disproportionately affect women.

Keywords: fear; learning; mouse; neuroscience; peptide.

Figure 1.. Partially reinforced fear drives a hyperarousal phenotype in males.

(A) Schematic of behavior chambers and tone/shock presentations for fully- and partially reinforced fear conditioning and recall. (B) Freezing during baseline and 0–28 s of tone presentations in Ctrl (two-way RM ANOVA, main effect of tone: F(3.137, 69.01)=1.449 ns, sex: F(1, 22)=1.717 ns, tone × sex: F(6, 132)=0.2104 ns), Full (two-way RM ANOVA, main effect of tone: F(2.046, 45.01)=45.19, p<.0001, sex: F(1, 22)=0.4527 ns, tone × sex: F(3, 66)=1.282 ns), and Part (two-way RM ANOVA, main effect of tone: F(3.218, 70.79)=70.73, p<0.0001, sex: F(1, 22)=0.3841 ns, tone × sex: F(6, 132)=0.4960 ns) fear. (C) Freezing for all tones from 0 to 28 s during fear conditioning (two-way ANOVA, main effect of fear: F(2, 66)=53.06, p<0.0001, sex: F(1, 66)=0.4488 ns, fear × sex: F(2, 66)=0.4588 ns). (D) Full tone 3 and Part tone 4 freezing during fear conditioning (two-way ANOVA, main effect of fear: F(1, 44)=0.1043 ns, sex: F(1, 44)=0.08787 ns, fear × sex: F(1, 44)=0.3009 ns). (E) Part fear tone 3, 4, 5 freezing during conditioning (two-way RM ANOVA, main effect of tone: F(2, 44)=6.578, p<0.01, sex: F(1, 22)=0.5318 ns, tone × sex: F(2, 44)=0.3213 ns), (F) Freezing during recall baseline and 0–28 s of tone presentations in Ctrl (two-way RM ANOVA, main effect of tone: F(3.394, 74.66)=1.449 ns, sex: F(1, 22)=1.261 ns, tone × sex: F(6, 132)=0.6956 ns), Full (two-way RM ANOVA, main effect of tone: F(4.980, 109.6)=12.19, p<0.0001, sex: F(1,22)=0.1797 ns, tone × sex: F(6, 132)=1.274 ns), and Part (two-way RM ANOVA, main effect of tone: F(4.754, 104.6)=14.43, p<0.0001, sex: F(1, 22)=0.0483 ns, tone × sex: F(6, 132)=1.065 ns) fear. (G) Freezing during baseline period in recall Full and Part fear (two-way ANOVA, main effect of fear: F(1, 43)=2.732 ns, sex: F(1, 43)=1.876 ns, fear × sex: F(1, 43)=1.776 ns). One male Part fear mice excluded as outlier. (H) Freezing during recall first tone exposure (two-way ANOVA, main effect of fear: F(1, 43)=1.228 ns, sex: F(1, 43)=0.0021 ns, fear × sex: F(1, 43)=1.687 ns) 1 female Part fear mouse excluded as outlier. (I) Average freezing for all tone presentations in Recall Full and Part Fear (two-way ANOVA, main effect of fear: F(1, 43)=7.401 p<.01, sex: F(1,43)=0.7413 ns, fear × sex: F(1, 43)=0.5288 ns) one female Part fear mouse excluded as outlier.(J) Behavioral timeline for Novelty-Induced-Suppression-of-Feeding (NSF) anxiety-like and Acoustic Startle arousal behavioral tests after fear conditioning. (K) Freezing for all tones from 0 to 28 s during fear conditioning (two-way ANOVA, main effect of fear: F(2, 54)=57.10, p<0.0001, sex: F(1, 54)=0.7934 ns, fear × sex: F(2, 54)=0.3522 ns). (L) Full tone 3 and Part tone 4 freezing during fear conditioning (two-way ANOVA, main effect of fear: F(1, 36)=2.236 ns, sex: F(1, 36)=0.5574 ns, fear × sex: F(1, 36)=0.3593 ns). (M) Part fear tone 3, 4, 5 freezing during conditioning (two-way RM ANOVA, main effect of tone: F(2, 36)=14.53, p<0.0001, sex: F(1, 18)=0.2412 ns, tone × sex: F(2, 36)=1.541 ns), (N) Latency to feed in the novelty-induced suppression of feeding (NSF). (O) Food consumed during NSF refeed test. (P) Acoustic startle response in males (two-way RM ANOVA, main effect of fear F(2, 20)=11.61, p=0.0005, main effect of stim intensity: F(1.808, 36.15)=166.9 p<0.0001, interaction: F(4,40)=6.096, p=0.0006, Tukey post-hoc: 120 dB Full v. Part, p=0.0395). One Ctrl 90 dB male excluded as outlier. (Q) Acoustic startle response in females (two-way RM ANOVA, main effect of fear F(2, 33)=1.869 ns, main effect of stim intensity F(1.7452, 57.57)=223.1, p<0.0001). n=8–12/group. (R) 120 dB acoustic startle response in males and females (two-way RM ANOVA, main effect of fear F(2, 53)=10.01, p<0.001, main effect of sex F(1, 53)=9.951, p<0.01, fear × sex: F(2, 53)=4.623, p<0.05: Tukey post-hoc Males Ctrl vs. Part p<0.0001, Males Full vs. Part p<0.01, Part Males vs. Females p<0.0001) 1 Male 90 dB Ctrl mice excluded as outlier. n=8–12/group. All data are shown as mean + SEM. *=p<0.05, **=p<0.01, ***=p<0.001, ****=p<0.0001.

Figure 2.. Bed nucleus of the stria terminalis (BNST) is dynamically engaged during fear conditioning and recall.

(A) Surgical schematic for fiber photometry experiments. (B, C) Schematic of tone and shock presentations during fear conditioning and recall. (D) BNST representative images showing fiber placement. (E) Fear Conditioning traces in males. (F) BNST response to tone onset (0–2 s) for all tones (two-way ANOVA, main effect of fear: F(1, 30)=5.834, p<0.05, sex: F(1, 30)=10.56, p<0.01, fear × sex: F(1, 30)=0.5614 ns; Bonferroni’s post-hoc test: Part Males vs Females p=0.0105). (G) Comparison of Full tone 3 and Part tone 4 BNST response (two-way ANOVA, main effect of fear: F(1, 30)=0.1296 ns, sex: F(1, 30)=4.216, p<0.05, fear × sex: F(1, 30)=0.8026 ns). (H) Fear Conditioning traces in females. (I) Shock response in Full and Part fear conditioning averaged from t=28–30 across all shock trials (two-way ANOVA, main effect of fear: F(1, 30)=0.7701 ns, sex: F(1, 30)=0.4407 ns, fear × sex: F(1, 30)=14.11, p<0.001, Bonferroni’s post-hoc: Males Full vs Part p=0.0068, Full Males vs Females p=0.0121). (J) Final shock response in Full and Part fear conditioning averaged from t=28–30 (two-way ANOVA, main effect of fear: F(1, 30)=3.142 ns, p=0.09, sex: F(1, 30)=0.2762 ns, fear × sex: F(1, 30)=5.682, p<0.05, Bonferroni’s post-hoc: Males Full vs Part p=0.0155). (K) BNST Tone response averaged from t=0–2 in Part fear across tone 3, 4, and 5 (two-way RM ANOVA, main effect of tone: F(2, 34)=0.1031 ns, sex: F(1, 17)=12.05, p<0.01, tone × sex: F(2, 34)=7.572, p<0.01, Tukey post-hoc: Males tone 4 vs 5, p=0.0147, Females tone 4 vs 5, p=0.0423, Tone 3 Males vs Females, p=0.0061, Tone 5 Males and Females, p<0.0001). (L) Shock response in Part fear averaged from t=28–30 across all shock trials (two-way RM ANOVA, main effect of tone: F(2, 34)=0.7578 ns, sex: F(1, 17)=3.788 ns, p=0.07, tone × sex: F(2, 34)=0.3663 ns). (M) Shock omission response in Part fear averaged from t=28–30 across no shock trials (two-way RM ANOVA, main effect of tone: F(2, 34)=0.8610 ns, sex: F(1, 17)=2.054 ns, tone × sex: F(2, 34)=1.689 ns). (N) Fear recall traces in males. (O) Fear recall traces in females. (P) BNST Tone response averaged from t=0–2 for all tones in Full and Part fear recall (two-way ANOVA, main effect of fear: F(1, 30)=0.02371 ns, sex: F(1, 30)=0.001158 ns, fear × sex: F(1, 30)=0.01792 ns). (Q) Area under the curve (AUC) during tone presentations from t=0–30 (two-way ANOVA, main effect of fear: F(1, 30)=0.7822 ns, sex: F(1, 30)=7.444, p<0.05, fear × sex: F(1, 30)=2.036 ns). n=7–10/group. All data are shown as mean + SEM. *=p<0.05, **=p<0.01, ***=p<0.001, ****=p<0.0001.

Figure 2—figure supplement 1.. Bed nucleus of the stria terminalis (BNST) activity negatively correlates with freezing.

(A) Schematic of viral infection. (B) Representative timeline showing freeze and mobile spiking of BNST. (C) Freezing during conditioning in males (D) Freezing during recall in males. (E) BNST spiking frequency during mobile/freeze mobility states in Full and Part fear conditioning males (two-way RM ANOVA, main effect of fear: F(1, 15)=2.295 ns, mobility: F(1, 15)=72.38, p<0.0001, fear × mobility: F(1, 15)=5.358, p<0.05, Bonferroni’s post-hoc: Full Mobile vs Freeze, p=0.0008, Part Mobile vs Freeze, p<0.0001, Mobile Full vs Part, p=0.0287). (F) BNST spiking frequency during recall mobile/freeze mobility states in males (two-way RM ANOVA, main effect of fear: F(1,15)=0.6208 ns, mobility: F(1, 15)=61.20, p<0.0001, fear × mobility: F(1, 15)=1.427 ns). (G) Freezing during conditioning in females (H) Freezing during recall in females. (I) BNST spiking frequency during mobile/freeze mobility states in Full and Part fear conditioning females (two-way RM ANOVA, main effect of fear: F(1,15)=0.3466 ns, mobility: F(1, 15)=49.44, p<0.0001, fear × mobility: F(1, 15)=0.0303 ns). (J) BNST spiking frequency during recall mobile/freeze mobility states in females (two-way RM ANOVA, main effect of fear: F(1,15)=0.0057 ns, mobility: F(1, 15)=17.79, p<0.001, fear × mobility: F(1, 15)=6.095, p<0.05, Bonferroni’s post-hoc: Part Mobile vs Freeze, p=0.0004).

Figure 3.. Corticotropin-releasing factor (CRF) knockdown in the bed nucleus of the stria terminalis (BNST) potentiates fear learning and recall in females.

(A) Surgical schematic and timeline of CRF knockdown experiments. (B, C) Representative image of Cre and GFP virus expression in BNST. (D) Males percent freezing to all tone exposures during fear conditioning (two-way ANOVA, main effect of fear: F(1, 36)=7.871, p<0.01, virus: F(1, 36)=0.1492 ns, fear × virus: F(1, 36)=0.4452 ns). (E) Comparison of males Full tone 3 and Part tone 4 freezing response (two-way ANOVA, main effect of fear: F(1, 36)=0.04672 ns, virus: F(1, 36)=0.004876 ns, fear × virus: F(1, 36)=0.3099 ns). (F) Males freezing response in Part fear across tone 3, 4, and 5 (two-way RM ANOVA, main effect of tone: F(1.899, 36.08)=6.882, p<0.01, virus: F(1, 19)=0.0044 ns, tone × virus: F(2, 38)=0.5234 ns). (G) Freezing during baseline period in males Full and Part fear recall. (H) Freezing during first tone presentation in males Full and Part fear recall. (I) Females percent freezing to all tone exposures during fear conditioning (two-way ANOVA, main effect of fear: F(1, 32)=16.71, p<0.001, virus: F(1, 32)=1.393 ns, fear × virus: F(1, 32)=0.4367 ns). (J) Comparison of females Full tone 3 and Part tone 4 freezing response (two-way ANOVA, main effect of fear: F(1, 32)=5.265, p<0.05, virus: F(1, 32)=0.8753 ns, fear × virus: F(1, 32)=0.3144 ns). (K) Females freezing response in Part fear across tone 3, 4, and 5 (two-way RM ANOVA, main effect of tone: F(1.596, 25.54)=3.840, p<0.05, virus: F(1, 16)=5.768, p<0.05, tone × virus: F(2, 32)=0.4641 ns). (L) Freezing during baseline in females Full and Part fear recall (two-way ANOVA, main effect of fear: F(1, 32)=0.3949 ns, virus: F(1, 32)=8.591, p<0.01, fear × virus: F(1, 32)=4.283, p<0.05, Bonferroni’s post-hoc: Part GFP vs Cre p=0.0025). (M) Freezing during first tone presentation in females Full and Part fear recall (two-way ANOVA, main effect of fear: F(1, 32)=1.628 ns, virus: F(1, 32)=4.545, p<0.05, fear × virus: F(1, 32)=1.886 ns). All data are shown as mean + SEM. *=p<0.05, **=p<0.01, ***=p<0.001, ****=p<0.0001.000.

Figure 4.. Corticotropin-releasing factor (CRF) knockdown in the bed nucleus of the stria terminalis (BNST) does not modulate anxiety-like or vigilance behaviors after fear.

(A) Surgical schematic and behavioral timeline. (B) Latency to feed in the novelty-induced suppression of feeding (NSF) test in males (two-way ANOVA, main effect of fear: F(2, 46)=15.24, p<0.0001, main effect of virus: F(1, 46)=1.808 ns, fear × virus interaction F(2, 46)=2.972 ns, p=0.06). (C) Latency to feed in the novelty-induced suppression of feeding (NSF) test in females (two-way ANOVA, main effect of fear F(2, 39)=1.190, p<0.05, main effect of virus F(1, 39)=1.190 ns, fear × virus interaction F(2, 39)=0.1847 ns). (D) Amount consumed during refeed test in males. (E) Amount consumed during refeed test in females. (F) Acoustic startle response in Ctrl mice (three-way RM ANOVA, main effect of dB: F(1.101, 30.83)=27.40, p<0.0001, virus: F(1, 28)=.0298 ns, sex: F(1, 28)=0.0167 ns). (G) Acoustic startle response in Full fear mice (three-way RM ANOVA, main effect of dB: F(1.701, 49.33)=126.3, p<0.0001, virus: F(1, 29)=3.590 ns, p=0.07, sex: F(1, 29)=5.260, p0.05, dB × Sex F(2, 58)=5.717, p<0.01). (H) Acoustic startle response in Part fear mice (three-way RM ANOVA, main effect of dB: F(1.466, 41.05)=88.49, p<0.0001, virus: F(1, 28)=.0037 ns, sex: F(1, 28)=4.763, p<0.05, dB × Sex F(2, 56)=3.965, p<0.05). (I) Acoustic startle response at 120 dB in males (two-way ANOVA, main effect of main effect of fear: F(2, 46) = 12.96 p<0.0001, main effect of virus: F(1, 46)=1.131 ns, fear × virus interaction: F(2, 46)=1.666 ns). (J) Acoustic startle response at 120 dB in females (two-way ANOVA, main effect of main effect of fear: F(2, 39) = 2.046 ns, main effect of virus: F(1, 39)=0.0080 ns, fear × virus interaction: F(2, 39)=0.0592 ns). N=7–10/group. All data are shown as mean + SEM. *=p<0.05, **=p<0.01, ***=p<0.001, ****=p<0.0001.

Figure 5.. Corticotropin-releasing factor (CRF) neurons differentially encode fear recall in males and females.

(A) Viral strategy and experimental design schematic. (B) Example traces of different responses to foot shock. (C) Pie charts showing the proportion of cells excited and inhibited in response to the first foot shock in males and females, fear groups combined. (D) Event frequency during baseline period in fear conditioning baseline (two-way ANOVA, main effect of fear: F(1, 161)=0.00030 ns, sex: F(1, 161)=7.558, p<0.01, fear × sex: F(1, 161)=3.987, p<0.05). (E) Event frequency as a change from baseline during consolidation period at the end of fear conditioning (two-way ANOVA, main effect of fear: F(1, 161)=0.3744 ns, sex: F(1, 161)=18.91, p<0.0001, fear × sex: F(1, 161)=0.6552 ns). (F) Event frequency during all tone presentations in fear conditioning (two-way ANOVA, main effect of fear: F(1, 161)=0.1263 ns, sex: F(1, 161)=6.462, p<0.05, fear × sex: F(1, 161)=7.135, p<0.01, Bonferroni Post Hoc: Full, Males vs Females, p=0.0011). (G) Event frequency during final shock tone, Full fear tone 3 and Part fear tone 4 (two-way ANOVA, main effect of fear: F(1, 161)=0.2.638 ns, sex: F(1, 161)=2.060 ns, fear × sex: F(1, 161)=0.02164 ns). (H) final shock in Full and Part fear conditioning (two-way ANOVA, main effect of fear: F(1, 161)=1.377 ns, sex: F(1, 161)=0.9991 ns, fear × sex: F(1, 161)=1.125 ns). (I) Example traces during baseline and trial end in fear recall. (J) Full and Part fear recall baseline period event frequency (two-way ANOVA, main effect of fear: F(1, 162)=0.05488 ns, sex: F(1, 162)=0.2847 ns, fear × sex: F(1, 162)=0.0058 ns). (K) Full and Part fear recall trial end event frequency as a change from baseline (two-way ANOVA, main effect of fear: F(1, 162)=8.357, p<0.01, sex: F(1, 162)=4.143, p<0.05, fear × sex: F(1, 162)=0.1101 ns, Bonferroni’s post-hoc: Males Full vs Part, p=0.0384). (L) Event frequency during all tone presentations in recall (two-way ANOVA, main effect of fear: F(1, 162)=4.593, p<0.05, sex: F(1, 162)=3.202 ns, p=0.08, fear × sex: F(1, 162)=0.0013 ns). (M) Comparison of event frequency during first and last tone presentation in recall (three-way ANOVA, main effect of fear: F(1, 162)=10.85, p<0.01, sex: F(1, 162)=7.385, p<0.01, fear × sex: F(1, 162)=3.981, p<0.05, fear × sex × tone: F(1, 162)=5.916, p<0.05, Tukey’s post hoc: Males Part first tone vs Males Part last tone, p=0.0004, Males Part last tone vs Females Part last tone, p=0.0005). (N) Comparison of event frequency during the 10 s following the first and last tone presentation in recall (three-way ANOVA, main effect of sex: F(1, 162)=3.340 ns, p=0.07, fear × sex: F(1, 162)=20.62, p<0.0001, fear × sex × tone: F(1, 162)=4.334, p<0.05, Tukey’s post hoc: Males Full first tone vs Males Full last tone, p=0.0069, Males Full first tone vs Females Full first tone, p=0.0237, Females Full last tone vs Females Part first tone, p=.0253). For fear conditioning: Full Males: n=41 cells from 6 mice. Part Males: n=39 cells from 5 mice. Full females: n=33 cells from 6 mice. Part females: n=52 cells from 5 mice. For fear recall: Full Males: n=45 cells from 6 mice. Part Males: n=44 cells from 5 mice. Full females: n=33 cells from 6 mice. Part females: n=45 cells from 5 mice. All data are shown as mean + SEM. *=p<0.05, **=p<0.01, ***=p<0.001, ****=p<0.0001.

Figure 6.. Sex differences in BNST^CRF neuron encoding of Part fear conditioning.

(A) Experimental design schematic for 1-photon recordings. (B) Representative raw FOV. (C) Representative processed FOV. (D) Representative image showing GCaMP expression and GRIN lens placement. (E) Event frequency during partial fear tone 3, 4, and 5 presentations (two-way RM ANOVA, main effect of tone: F(2, 180)=7.330, p<0.001, sex: F(1, 90)=0.3467 ns, tone × sex: F(2, 180)=6.365, p<0.01, Bonferroni’s post-hoc: Females tone 3 vs 4, p<0.0001, Females tone 3 vs 5, p<0.0001). (F) Event frequency during the 10 s post-shock across all shock trials (two-way RM ANOVA, main effect of shock: F(2, 178)=5.386, p<0.01, sex: F(1, 89)=0.0044 ns, shock × sex: F(2, 178)=1.243 ns). (G) Event frequency during the 10 s post-omission across all omission trials (two-way RM ANOVA, main effect of omission: F(2, 178)=1.412 ns, sex: F(1, 89)=8.552, p<0.01, omission × sex: F(2, 178)=0.3291 ns). (H) Event frequency during the 10 s final post-shock or post-omission window during partial fear (two-way RM ANOVA, main effect of shock/omission: F(1, 89)=0.0209 ns, sex: F(1, 89)=3.174 ns p=0.08, shock/omission × sex: F(1, 89)=4.697, p<0.05, Bonferroni’s post-hoc: Omission Males vs Females, p=0.0135). (I) Representative traces showing a response to shock and shock-omission trials for partial fear only in males and females. All data are shown as mean + SEM. *=p<0.05, **=p<0.01, ***=p<0.001, ****=p<0.0001.