PMC

(A) Timeline for ELS and behavioral testing. Schematic of control and ELS conditions from PN days 4–11. Control animals remain in their home cage with ample bedding, while ELS dam and pups are transferred into the ELS cage.
(B) Schematic of the attentional set-shifting task (black arrows, top) in which animals perform a simple discrimination (SD), followed by a compound discrimination (CD), a rule shift (RS), and a rule reversal (RR). The schematic with red arrows (bottom) shows that half of the mice underwent testing in a counterbalanced order, with RR preceding RS. Green text indicates the rewarded stimulus.
(C) A significant difference in the number of trials to reach criterion between task phases (repeated measure [RM] ANOVA; F_(3,34) = 5.145, p = 0.002) and a significant interaction between task phase × group (F_(3,34) = 2.470, p = 0.014) were found. A significant effect of group was found specifically in RR learning (F_(3,34) = 5.106, p = 0.005). ELS females (n = 8) took significantly more trials to reach criterion than ELS males and control females (n = 10) (Bonferroni multiple comparisons post hoc test; p = 0.029 and p = 0.004, respectively).
(D) There was also a significant difference in the number of errors committed between task phases (RM ANOVA; F_(3,34) = 6.635, p < 0.001) and an interaction between group and task phase (F_(3,34) = 2.675, p = 0.008). A significant difference between groups was again only observed in RR learning (F_(3,34) = 8.413, p < 0.001), with ELS females committing significantly more errors than any other group (Bonferroni multiple comparisons post hoc test; ELS female [F] versus unhandled control (UHC) male [M], p = 0.038; ELS M, p = 0.001; and UHC F, p = 0.001).
(E) No significant differences in latency were observed among groups in any task phase. For all plots, data show means ± SEM. Significance is denoted as follows: # for main effects or interaction effects, and * for significant post hoc comparisons; *p < 0.05, **p < 0.005. In the case of a significant one-way ANOVA, F statistics are reported, but only the significant post hoc Bonferroni comparisons are shown.

Quantitative real-time PCR analysis of relative mRNA expression of several interneuron markers—parvalbumin (PV), calbindin (Calb1), calretinin (Calb2), and somatostatin (SST), as well as neuropeptide Y (NPY) and GABA-synthesizing enzyme, and glutamic acid decarboxylase-67 kDa (GAD67)—was performed relative to control gene β-actin in the orbitofrontal cortex (left panels) and medial prefrontal cortex (middle panels). Top panels show female data (A–C); bottom panels show male data (D–F).
(A) In the OFC, female ELS-exposed mice had a significant and selective reduction of PV (t_(1,11) = 2.282, p = 0.043) and GAD67 (t_(1,11) = 2.429, p = 0.003) in comparison with control females. No other interneurons were affected by stress.
(B) In the mPFC of female animals, no effects of ELS were observed on PV, GAD67, NPY, or any other interneuron except a slight reduction in calbindin expression in ELS females (t_(1,9) = 2.302, p = 0.047).
(C) Cell counts in the OFC show ELS females had a significant decrease in PV+ density in both the ventral orbitofrontal (VO) cortex (t_(1,14) = 3.480, p = 0.004) and lateral orbitofrontal (LO) cortex (t_(1,14) = 3.533, p = 0.003) regions of the OFC in comparison with control females.
(D–F) In males, we observe(D) a significant reduction in PV mRNA gene expression in the OFC (t_(1,8) = 2.451, p = 0.040) of ELS mice, (E) no change in gene expression in the mPFC, and (F) this effect was not associated with changes in PV cell density in this region. No other interneuron markers were affected, nor were any changes detected in the mPFC.
(G) PV expression in the prefrontal cortex (PFC) increases across development in all groups (F_(3,53) = 36.404, p < 0.000) but is significantly affected by ELS (treatment × day interaction; F_(6,50) = 2.725, p = 0.024). Specifically, by P21, ELS females showed a significant reduction in PV expression in comparison with control animals (F_(1,13) = 6.020, p = 0.029).
(H) Images from the OFC of adult control and ELS females show PV+ immunohistochemical labeling (images at 203 magnification; scale bar, 50 ~M ). For all plots, data show means ± SEM and are analyzed using two-tailed Student’s t tests or ANOVA. Significance is denoted as follows: # for main effects or interaction effects, and * for post hoc or pairwise comparisons; *p < 0.05, **p < 0.005.

(A) Surgical and behavioral timeline for animals that underwent optogenetic stimulation in the OFC or mPFC. Examples of testing and illumination protocols on days 3 and 4 are shown, where yellow indicates continuous photostimulation with 620-nm light.
(B) In slice preparation from PV-Cre NpHR^+/− mice, we demonstrate optogenetic suppression of stimulus current-induced spiking in a NpHR^+/−-expressing PV interneuron. For slice experiments, a 545 ± band-pass excitation filter on white light emitting diode (LED) light was used. The light power at the plane of focus was 28.2 μW/mm². A more in-depth analysis of the effects of optogenetic stimulation is presented in .
(C) The effects of continuous illumination (yellow bars) over the OFC during various test phases of the attentional set-shifting task were assessed in NpHR^−/− mice (shown in black, n = 4) and NpHR^+/− mice (shown in gray, n = 8). There is a significant interaction among light stimulation, genotype, and task phase (three-way ANOVA; F_(2,69) = 7.963, p = 0.001). Specifically, NpHR^+/− mice take significantly more trials to reach criterion (t_(3,19) = 5.053, p = 0.0002) than NpHR^−/− littermates under light stimulation during RR learning.
(D) Similarly, under continuous illumination over the OFC, NpHR^+/− mice commit significantly more errors during RR learning than NpHR^−/− animals (t_(3,19) = 3.541, p = 0.0065). A significant three-way interaction among light stimulation, genotype, and task phase on errors commit during the task was also reported (F_(2,69) = 4.208, p = 0.019).
(E) Postmortem c-Fos analysis revealed continuous optical inhibition of PV interneurons in the OFC leads to an increase in c-Fos expression in this region t_(1,10) = 3.108, p = 0.011). Data shown for lateral orbitofrontal (LO) cortex; ventral orbitofrontal (VO) expression was also significant t_(1,10) = 2.841, p = 0.018) and can be found in .
(F) Effects of continuous photostimulation with 570-nm light over the mPFC. A significant interaction among light stimulation, genotype, and trial phase was reported (three-way ANOVA; F_(2,69) = 3.232, p = 0.045). Specifically, NpHR^+/− animals take significantly more trials to meet criterion during the RS phase than NpHR^−/− mice t_(8,31) = 5.686, p < 0.0001). These effects were not present in any other task phase or in the absence of light.
(G) Optogenetic inhibition of mPFC interneurons also affected errors committed during the task, with a three-way interaction found among light stimulation, genotype, and task phase (F_(2,69) = 4.590, p = 0.013). NpHR^−/+ mice (n = 6) committed significantly more errors than NpHR^−/− mice (n = 5) (t_(8,31) = 5.374, p < 0.0001) committed selectively during the RS task phase. Again, effects were only observed in the presence of photostimulation.
(H) Postmortem c-Fos analysis revealed that optically silencing PV interneurons in the mPFC led to an increase in c-Fos expression (t_(1,11) = 3.216, p = 0.008). Data shown for the infralimbic cortex (IL); expression in the prelimbic cortex (PL) was also significant (t_(1,11) = 2.883, p = 0.015) and can be found in .
For all plots, data show means ± SEM. Significant main effects and main interactions are reported, but only significant Bonferroni comparisons are shown. For c-Fos expression, data are analyzed using two-tailed Student’s t tests. Significance is denoted as follows: *p < 0.05, ***p < 0.001