The dorsal noradrenergic bundle carries ascending projections from A6 neurons, while the ventral noradrenergic bundle conveys ascending projections from A1 and A2 neurons through the mfb[18]. The fiber distribution suggests the proximity between the bundle and the electrode. For analysis, six sections at the anterior-posterior (AP) levels of -1.08 mm, -1.44 mm, -2.04 mm, -2.76 mm, -3.72 mm, and − 4.36 mm from Bregma were chosen according to the atlas in the review paper from Nieuwenhuys and colleagues[32].
We identified that the noradrenergic fibers project from lateral quadrate of the mfb posteriorly towards superior quadrate of the mfb anteriorly (Fig. 1A). Specifically, at the − 4.36mm and − 3.72mm level, the fibers were distinctly localized in the lateral quadrant of the mfb. Moving to the − 2.76 mm, the level of common implantation site DBS electrode, fibers were primarily located in the medial quadrant of the mfb. At -2.04 mm, ventromedial localization. The same ventromedial localization was also seen at the − 1.44 mm level. At the most anterior section − 1.08 mm these fibers were predominantly located within both the lateral and medial quadrants of the mfb (Fig. 1B). In the FSL depressive model, lower amount of noradrenergic fibers was found through different levels compared with control (t = 9.067, df = 10, p < 0.0001) (Fig. 1C, D).
At the standard implantation site of DBS electrode, noradrenergic fibers were primarily localized near the fornix (Fig. 1A). Employing fornix as the landmark under a 100X confocal microscope, we observed that the thin noradrenergic fibers were unmyelinated running with large myelinated axons (Fig. 1E, F). This characteristic suggests that the majority of these slender monoamine fibers are unlikely to be directly stimulated by our DBS protocol at hundred microsecond pulse width range.
Key Insights into Strain-Specific Responses to Deep Brain Stimulation Parameters
The experimental timeline is depicted in Fig. 2A. Before surgical intervention, both FSL and SD rats underwent phenotyping using the FST. FSL rats demonstrated significantly higher total immobility times compared to their SD counterparts (t = 3.685, df = 15, p = 0.0022, Fig. 2D). AAV9-hsyn-NE2m was injected and optic fibers were implanted unilaterally in the PFC (FSL n = 5, SD n = 5) and NAC shell (FSL n = 6, SD n = 5) as illustrated in Fig. 2E. The DBS electrode was positioned on the ipsilateral side of the mfb. After four weeks adequate viral expression, we explored six distinct stimulation parameters, including frequency and pulse width, applied in a random order on different days. This also involved a clinically established parameter of 130 Hz frequency with a 100 µs pulse width. During each session, a 5-second mfb DBS was applied, followed by an observation of the NA spike. Subsequently, NA release over a 50-second span was measured and averaged across 20 repeated trials (Fig. 2B). The stimulation prompted a release of NA in both the PFC and NAC, as seen in Fig. 2F-Q. However, there were disparities across the different stimulation parameters and the experimental groups. Regardless of alterations in pulse width or frequency, the NA levels increased in a similar manner. The release started and plateaued within the range of the selected parameters in which 30 Hz with 100 µs pulse width, minimal or non-significant NA release was observed and NA release appeared to plateau at 130 Hz 250 µs (PFC) and 30 Hz 250 µs (NAC). Notably, FSL rats showed a delay in the peak in NA release by approximately 1–2 seconds compared to SD rats, potentially indicating divergent neuronal mechanisms in FSL rats.
In PFC, two-way ANOVAs did not indicate a statistically significant interaction between the time and animal strain. However, time had a statistically significant effect on NA release (30 Hz 100 µs, p = 0.0235, Fig. 2F; 30 Hz 250 µs, p = 0.0013, Fig. 2G; 30 Hz 350 µs, p = 0.0009, Fig. 2H; 130 Hz 100 µs, p = 0.003, Fig. 2I; 130 Hz 250 µs, p = 0.0029, Fig. 2J; 130 Hz 350 µs, p = 0.0017, Fig. 2K). Furthermore, animal strains did not have a statistically significant effect on NA release in PFC. In NAC, a different pattern of NA release was observed compared to the PFC. Specifically, FSL rats demonstrated a more pronounced increase in NA levels within the NAC. Two-way ANOVAs revealed a significant interaction between time and animal strain, but only at higher frequency or pulse width settings. (30 Hz 350 µs, F(1,9) = 6.188, p = 0.346, Fig. 2N; 130 Hz 250 µs, F(1,9) = 6.486, p = 0.0314, Fig. 2P; 130 Hz 250 µs, F(1,9) = 5.496, p = 0.0437, Fig. 2Q). The main effects analysis revealed that time exerted a significant influence on NA release, except when a 30 Hz/100 µs (Fig. 2L stimulation parameter was employed (30 Hz 250 µs, p = 0.0120, Fig. 2M; 30 Hz 350 µs, p = 0.0033, Fig. 2N; 130 Hz 100 µs, p = 0.0293, Fig. 2O; 130 Hz 250 µs, p = 0.0016, Fig. 2P; 130 Hz 350 µs, p = 0.0023, Fig. 2Q). Moreover, the simple main effects analysis also indicated that the animal strain had a significant impact on NA release in the NAC, specifically at higher frequencies or pulse widths (30 Hz 350 µs, p = 0.0360, Fig. 2N; 130 Hz 250 µs, p = 0.0361, Fig. 2P; 130 Hz 350 µs, p = 0.0432, Fig. 2Q). Post-hoc analyses showed that significant changes were observed in the FSL group (30 Hz 350 µs, p = 0.0020, Fig. 2N; 130 Hz 250 µs, p = 0.0011, Fig. 2P; 130 Hz 350 µs, p = 0.0018, Fig. 2Q), rather than the SD rats.
Ultrasonic vocalization, an index of the rodents' affective state, was registered simultaneously with fiber photometry recording throughout the experiment. Vocalizations within the 40 kHz to 60 kHz range representing a positive affective state were analyzed (Fig. 3A). The number and duration of calls between the FSL and SD rats were compared across four distinct time intervals: T0 (pre-DBS), T1 (during DBS), T2 (5-10s post-DBS), and T3 (averaged over the final 40s post-DBS). While no significant differences emerged between the SD and FSL rats during the T0 and T1 intervals (Fig. B, C, G, F), post-DBS periods (T2 and T3, Fig. 3D, E, H, I) in FSL rats gradually increase their positive vocalizations, distinguishing them significantly from SD rats in both number of the calls (T1, U = 8, p = 0.0465 ; T2, U = 3, p = 0.0065) and length of the calls (T1, U = 7, p = 0.038 ; T2, U = 3, p = 0.007). These findings provide valuable insights into the nuanced effects of deep brain stimulation parameters on NA release and behavioral outcomes across the experimental groups, and identifies possible mechanisms for further research aimed at tailoring DBS settings for optimized therapeutic outcomes.
mfb DBS activates A1, A2 noradrenergic cell groups and disinhibited the neurons in PFC and NAC
To elucidate the origins of noradrenergic sources, we assessed the c-Fos expression through DBH/c-Fos double staining of the noradrenergic cell groups A1, A2, and A6, which project to the forebrain from the brainstem (Fig. 4A, E, I). While the cell density of A1 and A2 neurons displayed no significant differences between SD and FSL rats (Fig. 4B, F), the A6 neurons in FSL rats exhibited a notably lower normalized total cell count compared to SD rats (t = 3.421, df = 24, p = 0.0022, Fig. 4J). In the unstimulated rats, the neurons in A1 in both FSLs and SDs (MSD = 17%, MFSL= 18.7%, Fig. 4C, D) and in A2 uniquely in FSLs (MFSL= 28.3%, Fig. 4H), were less active. The neurons in A6 in both groups (MSD = 36.3%, MFSL= 62%, Fig. 4K, L) and A2 in SDs (MSD = 50%, Fig. 4G) were more active. Two-way ANOVAs were performed to analyze the effect of stimulation and brain side on NA cell groups activation in SD and FSL rats after 20 times 5 second mfb DBS stimulation. In none of the comparisons implicating A1, A2, or A6 and FSLs or SDs, were statistically significant interactions between the clinical parameter stimulation and brain side. However, stimulation did have a statistically significant effect on A1 NA neurons’ activation (p = 0.0003, Fig. 4C) in SD rats, but this was not observed in either A2 or A6 (Fig. 4G, K). Furthermore, despite unilateral stimulation, in both SDs and FSLs, the NA neurons in the examined brain nuclei were similarly effected on both sides of the brain. In FSL rats, stimulation had statistically significant effect on A1 and A2 NA neurons’ activation (A1 and A2: p < 0.0001, Fig. 4D, H) but not on A6 (Fig. 4L).
In the PFC and NAC, the total density, baseline activated density and activation in stimulated and sham of PV interneuron were compared in SD and FSL rats. The activation was measured by colocalization of PV interneuon with c-Fos staining (Fig. 5A, H). Subsequentially, the total c-Fos density in stimulated and sham was also compared in SD and FSL rats. In PFC, SD rats had a significant more PV interneuron than FSL rats (t = 2.923, df = 15, p = 0.0105, Fig. 5B). However, FSL rats had more activated PV than SD rats in sham animal (t = 5.454, df = 4, p = 0.0055, Fig. 5C). Stimulation in FSL rats normalized PV interneuron activation to levels comparable to SD rats (Fig. 5D, E). Two-way ANOVAs were performed to analyze the effect of stimulation and brain side on PV interneuron activation in PFC of SD and FSL rats after 20 times 5 second mfb DBS stimulation. In both SDs and FSLs, the analysis did not indicate a statistically significant interaction between the stimulation and brain side. However, stimulation did have a statistically significant effect on PV interneuron activation (p < 0.0001, Fig. 5D) in FSL rats but not on the SD rats.
In NAC, FSLs had significantly higher total density compared to SDs (t = 3.958, df = 14, p = 0.0014, Fig. 5I) and density of activated PV interneurons (t = 7.725, df = 4, p = 0.0015, Fig. 5J). Two-way ANOVAs did not indicate a statistically significant interaction between the variables of stimulation and brain side. However, the stimulation did have a statistically significant effect on PV interneuron activation (p = 0.0002, Fig. 5J) in FSL rats but not on the SD rats (Fig. 5I). The Brain side did not have a statistically significant effect on PV interneuron activation in the PFC and NAC.
c-Fos positive neurons –an index of neuronal activation - were quantified in the areas of interest. Two-way ANOVAs did not indicate a statistically significant interaction between the variables of stimulation and brain side. The stimulation significantly increased c-Fos positive neurons in NAC of FSL rats (p = 0.0002, Fig. 5N) but not on the SD rats (Fig. 5M); nor where there any stimulation mediated differences in either of the groups in the PFC (Fig. 5F, G). Both sides of the brain were equally effected concerning PV interneuron activation in the NAC.