Bidirectional optogenetic modulation of prefrontal-hippocampal connectivity in pain-related working memory deficits

Abstract

Dysfunction of the prefrontal-hippocampal circuit has been identified as a leading cause to pain-related working-memory (WM) deficits. However, the underlying mechanisms remain poorly determined. To address this issue, we implanted multichannel arrays of electrodes in the prelimbic cortex (PL-mPFC), and in the dorsal hippocampal CA1 field (dCA1) to record the neural activity during the performance of a delayed non-match to sample (DNMS) task. The prefrontal-hippocampal connectivity was selectively modulated by bidirectional optogenetic inhibition or stimulation of local PL-mPFC glutamatergic calcium/calmodulin-dependent protein kinase-II alpha (CaMKIIα) expressing neurons during the DNMS task delay-period. The within-subject behavioral performance was assessed using a persistent neuropathic pain model - spared nerve injury (SNI). Our results showed that the induction of the neuropathic pain condition affects the interplay between PL-mPFC and dCA1 regions in a frequency-dependent manner, and that occurs particularly across theta oscillations while rats performed the task. In SNI-treated rats, this disruption was reversed by the selective optogenetic inhibition of PL-mPFC CaMKIIα-expressing neurons during the last portion of the delay-period, but without any significant effect on pain responses. Finally, we found that prefrontal-hippocampal theta connectivity is strictly associated with higher performance levels. Together, our findings suggest that PL-mPFC CaMKIIα-expressing neurons could be modulated by painful conditions and their activity may be critical for prefrontal-hippocampal connectivity during WM processing.

Figure 1

Expression of eNpHR3.0 and hChR2 in PL-mPFC CaMKIIα neurons, neuroelectrophysiological recording locations, functional optogenetic modulation effects in local neural activity, working-memory task and experimental timeline. (A) Expression of eNpHR3.0-eYFP (left side) and hChR2-mCherry (right site) in PL-mPFC brain area. Both opsins are expressed in prelimbic (PL), but not in infralimbic (IL) and anterior cingulate cortex (Cg1) regions of medial prefrontal cortex. (B) Panel showing an original magnification (x20) of an PL-mPFC eNpHR3.0-expressing neuron, and (C) an hChR2-expressing neuron. Blue dots represent the DAPI DNA-labeling. (D) The location of optogenetic viral particles infusion and optical fiber in the PL-mPFC, and experimental setup for optogenetic light modulation of local PL-mPFC activity and simultaneous LFPs signals acquisition in the PL and dorsal hippocampus (dCA1). (E) Example of the optogenetic inhibition effect in a PL-mPFC CaMKIIα-eNpHR3.0-expressing neuron, and (F) example of the optogenetic stimulation effect in a PL-mPFC CaMKIIα-hChR2-expressing neuron. Optogenetic selective inhibition implemented using an orange led light (continuous pulse at 5 mW@620 nm; led on - background orange window), and selective stimulation using a blue led light (pulse duration of 15 ms, frequency of 10 Hz, and a fix intensity of 5 mW@465 nm; led on - background blue window). (G) Diagram of delayed nonmatch-to-sample task (DNMS) used in this study. Each trial began with a single lever being exposed (sample phase). When the animal pressed the lever it retracted and the delay-period were initiated. At the end of this period, both levers were exposed (choice-phase) and the animal need to press the opposite lever selected during sample phase to obtain a reward pellet. (H) Timeline of the experimental protocol. (I) Learning curve, gain in performance during 10 training sessions using 1 s delay-period.

Figure 2

Optogenetic modulation of PL-mPFC CaMKIIα-expressing neurons did not influence pain responses, but induced changes in working-memory performance. (A) Effects of contralateral optogenetic inhibition (left panel) and stimulation (right panel) of PL-mPFC CaMKIIα-expressing neurons in mechanical sensivity threshold. (B) Illustration of light modulation protocols applied during the delay-period. (C) Behavioral performance of eNpHR3.0-expressing rats using a delay-period challenge of 1 (left panel), 3 (middle panel), and 6 s (right panel). (D) Behavioral performance of hChR2-expressing rats using a delay-period challenge of 1 (left panel), 3 (middle panel), and 6 s (right panel). eNpHR3.0-expressing rats: sham n = 5, and SNI n = 5; and hChR2-expressing rats: sham n = 5, and SNI n = 4. Comparisons between experimental groups and light delivery protocols are based on Mann-Whitney test (for single comparisons) or Kruskal-Wallis test (for multiple comparisons) followed by post hoc Dunn’s test. Values are presented as mean ± SEM. *p < 0.05, and **p < 0.01.

Figure 3

Optogenetic modulation of PL-mPFC CaMKIIα-expressing neurons induced changes in theta power oscillations during working-memory delay-period. Effects of contralateral optogenetic inhibition of eNpHR3.0-expressing rats in (A) PL-mPFC and (B) dCA1 LFP theta (θ, 4–9 Hz) frequency-band power activity during 3 (left panels) and 6 s (right panels) DNMS delay-period challenges. Effects of contralateral optogenetic stimulation of hChR2-expressing rats in (C) PL-mPFC and (D) dCA1 LFP theta frequency-band power activity during 3 (left panels) and 6 s (right panels) DNMS delay-period challenges. Data were calculated for each entire recording session independently of correct and incorrect trials. Box-and-whiskers plots are based in the mean and minimum/maximum values. eNpHR3.0-expressing rats: sham n = 5, and SNI n = 5; and hChR2-expressing rats: sham n = 5, and SNI n = 4. Comparisons between experimental groups and light delivery protocols are based on Kruskal-Wallis test followed by post hoc Dunn’s test. Values are presented as mean ± SEM. *p < 0.05, and **p < 0.01.

Figure 4

Optogenetic modulation of PL-mPFC CaMKIIα-expressing neurons induced changes in prefrontal-hippocampal theta coherence and phase-coherence during working-memory delay-period. (A) Effects of contralateral optogenetic inhibition of eNpHR3.0-expressing rats in prefrontal-hippocampal LFP signals theta (θ, 4–9 Hz) quadratic coherence activity during 3 (left panel) and 6 s (right panel) DNMS delay-period challenges. (B) Effects of contralateral optogenetic stimulation of hChR2-expressing rats in prefrontal-hippocampal LFP signals theta quadratic coherence activity during 3 (left panel) and 6 s (right panel) delay-period challenges. (C) Effects of contralateral optogenetic inhibition of eNpHR3.0-expressing rats in prefrontal-hippocampal LFP signals theta phase-coherence activity during 3 (left panel) and 6 s (right panel) delay-period challenges. (D) Effects of contralateral optogenetic stimulation of hChR2-expressing rats in prefrontal-hippocampal LFP signals theta phase-coherence activity during 3 (left panel) and 6 s (right panel) delay-period challenges. Data were calculated for each entire recording session independently of correct and incorrect trials. Box-and-whiskers plots are based in the mean and minimum/maximum values. eNpHR3.0-expressing rats: sham n = 5, and SNI n = 5; and hChR2-expressing rats: sham n = 5, and SNI n = 4. Comparisons between experimental groups and light delivery protocols are based on Kruskal-Wallis test followed by post hoc Dunn’s test. Values are presented as mean ± SEM. *p < 0.05, and **p < 0.01.

Figure 5

The optogenetic inhibition of PL-mPFC CaMKIIα-expressing neurons alters significantly the prefrontal-hippocampal connectivity in SNI-treated rats during the 3 s delay-period challenge. The bidirectional analysis of prefrontal-hippocampal connectivity (PDC activity) showed that changes in circuitry connectivity occurred mainly at the theta (θ, 4–9 Hz) frequency-band during DNMS task 3 s delay-period challenge. The selective optogenetic inhibition of PL-mPFC CaMKIIα-eNpHR3.0-expressing neurons increased the prefrontal-hippocampal theta connectivity in SNI-treated rats, and decreased from the hippocampus to the prefrontal cortex in sham-treats rats. In the case of activation of PL-mPFC CaMKIIα-hChR2-expressing neurons, both experimental groups shared a decreased of their prefrontal-hippocampal connectivity. The effects of optogenetic inhibition of eNpHR3.0-expressing rats in PDC activity: (A) without inhibition (led off); and (B) with light inhibition applied during the whole delay-period. The effects of optogenetic activation of hChR2-expressing rats in PDC activity: (C) without activation (led off); and (D) with light activation applied during the whole delay-period. Top panels indicate PDC activity from PL-mPFC to dCA1 direction, whereas bottom panels indicate PDC activity from dCA1 to PL-mPFC. The respective right panels indicate PDC activity from 1–100 Hz (step resolution of 1 Hz) for each experimental groups and animal. Data were calculated for each entire recording session independently of correct and incorrect trials. eNpHR3.0-expressing rats: sham n = 5, and SNI n = 5; and hChR2-expressing rats: sham n = 5, and SNI n = 4. Comparisons between experimental groups across light delivery protocols are based on two-way ANOVA (F1: experimental groups X F2: frequency-bands) test followed by post hoc Bonferroni test. Values are presented as mean ± SEM. *p < 0.05, and ***p < 0.001.

Figure 6

The optogenetic inhibition of the PL-mPFC CaMKIIα-expressing neurons during the late phase of the 6 s delay-period challenge increased the prefrontal-hippocampal connectivity in both experimental groups. The optogenetic inhibition of PL-mPFC CaMKIIα-eNpHR3.0-expressing neurons decreased the prefrontal-hippocampal theta connectivity in both experimental groups during the whole delay-period, but not when applied during the late portion of the delay-period. In the case of activation of PL-mPFC CaMKIIα-hChR2-expressing neurons, both experimental groups shared no significant differences in their prefrontal-hippocampal connectivity (both neuromodulation protocols). The effects of optogenetic inhibition of eNpHR3.0-expressing rats in PDC activity: (A) without inhibition (led off); (B) with light inhibition applied during the whole delay-period; and (C) with light inhibition during the late phase of the delay-period. The effects of optogenetic activation of hChR2-expressing rats in PDC activity: (D) without activation (led off); (E) with light activation applied during the whole delay-period; and (F) with light activation during the late phase of the delay-period. Top panels indicate PDC activity from PL-mPFC to dCA1 direction, whereas bottom panels indicate PDC activity from dCA1 to PL-mPFC. eNpHR3.0-expressing rats: sham n = 5, and SNI n = 5; and hChR2-expressing rats: sham n = 5, and SNI n = 4. Comparisons between experimental groups across light delivery protocols are based on two-way ANOVA (F1: experimental groups X F2: frequency-bands) test followed by post hoc Bonferroni test. Values are presented as mean ± SEM. *p < 0.05, and ***p < 0.001.

Figure 7

Prefrontal-hippocampal theta connectivity is a good predictor for correct working-memory performance. Perievent raster plots representing the prefrontal-hippocampal theta (θ, 4–9 Hz) connectivity during correct (A) 3 s DNMS delay-period; and (B) 6 s DNMS delay-period challenge) and incorrect trials (C) 3 s DNMS delay-period; and (D) 6 s DNMS delay-period challenge). Correct trials are characterized by a transient bidirectional prefrontal-hippocampal connectivity enhancement at theta frequency-band before level press, which decreased after lever press. Incorrect trials are characterized by an opposite response. Each left panel indicate the response to sham-treated rats, whereas each right panel indicate the response to SNI-treated rats. Top panels indicate PDC activity from PL-mPFC to dCA1, whereas bottom panels from dCA1 to PL-mPFC. Frequency-time function computed with a resolution of 1 Hz, and centered at free-choice lever press (time = 0 s; vertical gray line). The averaged bidirectional prefrontal-hippocampal theta connectivity, before and after free-choice lever press, for 3 (E) from PL-mPFC to dCA1; and (F) from dCA1 to PL-mPFC) and 6 s (G) from PL-mPFC to dCA1; and (H) from dCA1 to PL-mPFC) DNMS delay-period challenges. Left panels indicate correct trials, and right panels indicate incorrect trials. (I) Fitting activity prediction model accuracy behavior. Model computed using randomly 200 trials. This model was used to classify performed trials according to PDC activity oscillation after free-choice lever press. If PDC_d < 0 as a correct trial, otherwise, if PDC_d≥0 as an incorrect trial. Comparisons between experimental groups and lever press responses are based on Kruskal-Wallis test followed by post hoc Dunn’s test. Values are presented as mean ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001.