Histochemical Characterization of the Dorsal Raphe-Periaqueductal Grey Dopamine Transporter Neurons Projecting to the Extended Amygdala

Abstract The dorsal raphe (DR) nucleus contains many tyrosine hydroxylase (TH)-positive neurons which are regarded as dopaminergic (DA) neurons. These DA neurons in the DR and periaqueductal gray (PAG) region (DADR-PAG neurons) are a subgroup of the A10 cluster, which is known to be heterogeneous. This DA population projects to the central nucleus of the amygdala (CeA) and the bed nucleus of the stria terminalis (BNST) and has been reported to modulate various affective behaviors. To characterize, the histochemical features of DADR-PAG neurons projecting to the CeA and BNST in mice, the current study combined retrograde labeling with Fluoro-Gold (FG) and histological techniques, focusing on TH, dopamine transporter (DAT), vasoactive intestinal peptide (VIP), and vesicular glutamate transporter 2 (VGlut2). To identify putative DA neurons, DAT-Cre::Ai14 mice were used. It was observed that DATDR-PAG neurons consisted of the following two subpopulations: TH+/VIP– and TH–/VIP+ neurons. The DAT+/TH–/VIP+ subpopulation would be non-DA noncanonical DAT neurons. Anterograde labeling of DATDR-PAG neurons with AAV in DAT-Cre mice revealed that the fibers exclusively innervated the lateral part of the CeA and the oval nucleus of the BNST. Retrograde labeling with FG injections into the CeA or BNST revealed that the two subpopulations similarly innervated these regions. Furthermore, using VGlut2-Cre::Ai14 mice, it was turned out that the TH–/VIP+ subpopulations innervating both CeA and BNST were VGlut2-positive neurons. These two subpopulations of DATDR-PAG neurons, TH+/VIP– and TH–/VIP+, might differentially interfere with the extended amygdala, thereby modulating affective behaviors.


Introduction
The dorsal raphe nucleus (DR) has been recognized as a typical serotonergic (5-HT) nucleus in mammals (Takeuchi et al., 1982a, b;Ishimura et al., 1988). To date, diverse lines of studies have focused on the physiological roles of 5-HT neurons (Hornung, 2003;Muller and Jacobs, 2010;Ren et al., 2018). However, the DR not only consists of 5-HT neurons but contains also many tyrosine hydroxylase (TH) positive neurons in rodents (Trulson et al., 1985;Descarries et al., 1986;Hasue and Shammah-Lagnado, 2002;Hioki et al., 2010;Poulin et al., 2014) and primates (Arsenault et al., 1988;Charara and Parent, 1998). Recently, the functional importance of dopaminergic (DA) neurons on affective behaviors in the DR and ventral periaqueductal gray (PAG) regions has been reported in multiple studies in mice. DA neurons in the DR-PAG region (DA DR-PAG ) influence pain sensation (Li et al., 2016;Yu et al., 2021), modulate social interactions (Matthews et al., 2016), accelerate fear learning (Groessl et al., 2018), respond to salient stimuli (Cho et al., 2017(Cho et al., , 2021, and control memory related to reward incentives (Lin et al., 2020(Lin et al., , 2021. These studies pointed to the bed nucleus of the stria terminalis (BNST) and the central nucleus of the amygdala (CeA), which are known to play significant roles in affective behaviors, as downstream targets of DA DR-PAG neurons. Projections from TH neurons in the DR to these regions have been confirmed using a retrograde tracer (Hasue and Shammah-Lagnado, 2002).
However, a further complication is that the DA transporter (DAT)-positive neurons in the DR-PAG region (DAT DR-PAG neurons), which are expected to be DA neurons, appear to have an array of diversity in colocalization with other neuron markers. At least mRNAs encoding vasoactive intestinal peptide (VIP), vesicular glutamate transporter 2 (VGlut2), calbindin, cholecystokinin, and TH are known to be expressed in these neurons (Poulin et al., 2014;La Manno et al., 2016;Hook et al., 2018;Tiklová et al., 2019; for review, see Poulin et al., 2020). This heterogeneity raises the possibility that each population of DAT DR-PAG neurons plays a unique role in affective behaviors, while fine features of connections from the heterogeneous DA DR-PAG neurons to the CeA and BNST remain unclear. Also, the presence of mRNA does not necessarily ensure that the encoded protein is expressed sufficiently to a physiologically relevant extent, suggesting that the confirmation of actual protein expression may also be important in evaluating heterogeneity.
Therefore, in the present study, we aimed to clarify the potential heterogeneity of DA DR-PAG neurons projecting to the CeA and BNST. We applied immunohistochemistry and tracer methods focusing on DAT, TH, VIP, and VGlut2 in DAT-Cre mice (Bäckman et al., 2006), in which expression of Cre-recombinase is more restricted to DA neurons than in TH-Cre mice (Savitt et al., 2005;Lammel et al., 2015;Stuber et al., 2015). The present experiments thus showed that the majority of DAT neurons projecting onto the CeA and BNST consist of two different subpopulations. These subpopulations of DAT DR-PAG neurons might play different roles in the regulation of affective behaviors.

Materials and Methods
Experiments were performed in accordance with the guiding principles of the Physiologic Society of Japan and were approved by the Animal Care Committee of Kanazawa Medical University and University of Toyama.

Immunohistochemistry
After surgery, the animals were maintained in home cages. The survival period was 10 d for FG or 6-OHDA-injected mice and 14-21 d for AAV-injected mice. The subjects were deeply anesthetized with isoflurane and were perfused transcardially with 4% paraformaldehyde in 0.1 M PB (pH 7.4). Brains were dissected out and immersed in the same fixative overnight and were cryoprotected with 30% sucrose in 0.1 M PB (pH 7.4) for 2 d. Then, frozen coronal sections were cut to a thickness of 40 mm using a microtome (Yamato Koki, Saitama, Japan) equipped with a freezing stage. Every sixth section was used for histology. For fluorescent immunohistochemistry, sections were incubated overnight in some combinations of rabbit anti-FG (1:2000; AB153-I, Merck-Millipore, MA, USA), rat anti-RFP (1:1000; 5F8, Chromotek, Planegg-Martinsried, Germany), rat anti-DAT (1:500; MAB369, Merck-Millipore), mouse anti-TH (1:1000, MAB318, Merck-Millipore), guinea pig anti-VIP (1:200; Hioki et al., 2018), and rat anti-GFP antibodies (1:500; 04404, Nacalai Tesque, Kyoto, Japan) diluted in incubation buffer, which was composed of 1% normal donkey serum, PBS, 0.3% Triton X-100, and 0.02% sodium azide. On the second day, sections were washed and incubated for 2 h with secondary antibodies raised from donkey conjugated with Alexa Fluor 488, 594, 647, or Cy5 (1:200; Jackson ImmunoResearch, PA, USA). In some cases, propidium iodide (PI; 1:2000) was added in secondary antibody solution to label the Nissl substance. For FG signals, tyramide signal amplification (TSA) was applied. For the TSA reaction, a biotinylated anti-rabbit secondary antibody was used instead of a fluorescent secondary antibody. The sections were then incubated with avidin-biotinylated horseradish peroxidase complex (1:50, ABC Elite, Vector Laboratories, CA, USA) diluted in PBS with 0.3% Triton X-100. The sections were processed with biotinylated-TSA Kuramoto et al., 2009), and then incubated with Alexa Fluor 488-conjugated streptavidin (1:400, S11223, Invitrogen, MA, USA). Sections were mounted on coated slides, air-dried, rehydrated, and coverslipped with ProLong Glass (Thermo Fisher Scientific, MA, USA). The combinations of primary and secondary antibodies or TSA in each experiment are shown in Table 1. For the brightfield immunohistochemistry, sections were incubated overnight with one of the primary antibodies, washed, and incubated for 2 h with biotinylated secondary antibodies that were raised from donkey (1:200; Jackson ImmunoResearch). Then, sections were washed and incubated for 2 h with ABC (Vector). After this step, the bound peroxidase was visualized using a nickel diaminobenzidine (DAB) reaction. Sections were mounted on coated slides, air-dried, dehydrated, and coverslipped with Entellan (Merck-Millipore).
Unless otherwise noted, fluorescence images were acquired using a laser scanning confocal microscope (LSM710, Carl Zeiss Microimaging, Jena, Germany). Alexa Fluor 488 was excited by a 488-nm Ar laser, and the emitted fluorescence was filtered with a 493-to 556-nm bandpass filter. Alexa Fluor 594 was excited by a 561 nm Diode Pumped Solid State laser, and the emitted fluorescence was filtered with a 589-to 628-nm bandpass filter. Alexa Fluor 647 and Cy5 were excited by a 633-nm He-Ne laser and emitted fluorescence was filtered with a 643to 759-nm low-pass filter. Images of each dye were taken sequentially to avoid bleed-through artifacts. The fluorescence intensity levels were adjusted using FIJI (ImageJ, NIH, MD, USA). Brightfield images were acquired using a microscope (BZ-9000, Keyence, Osaka, Japan).

The combination of fluorescent in situ hybridization and immunohistochemistry
Digoxigenin (DIG)-labeled antisense riboprobes were made from cDNAs of mouse DAT (nucleotides of 2-210, GenBank accession number: NM_010020.3). The specificity of the riboprobe was determined by the absence of staining in sections reacted with the sense riboprobe.

Data analysis
FIJI (ImageJ, NIH) was used to quantify the fiber signal intensity and normalized fluorescence intensity. The fiber signal intensity determined here was the subtraction of the background signal intensities from the signal intensity in each region. The normalized fluorescence intensity determined here was the fluorescence intensity in the lateral part of CeA (CeL) or the oval nucleus of BNST (BNST OV) divided by the fluorescence intensity in the striatum in the same section. These intensities for each region were sampled from the right hemisphere of each subject. Data are expressed as means 6 SEM. Permutation t tests followed by Holm's correction were used for the multiple comparisons and permutation t tests were used to compare averages. Shapiro-Wilk test was used to assess the normality of distributions. Statistical significance was set at p , 0.05. Mean, p-value, confidence interval (CI), and some panels were calculated and generated by using the web application of Estimation stats (Ho et al., 2019). Information for the primary statistical analysis is provided in Table 2.

DAT neurons in the DR-PAG region consist of heterogeneous subpopulations
First, we examined the expression of DAT in the DR-PAG region (DAT DR-PAG ), depending on the expression of tdTomato of DAT-cre::Ai14 mice. As shown in Figure 1A, the presence of DAT-positive cell bodies was confirmed in the DR-PAG region from the ventral PAG to the linear nucleus. This is consistent with previous reports that described the presence of DA neurons in this region (Li et al., 2016;Matthews et al., 2016;Cho et al., 2017;Groessl et al., 2018). With anti-TH and anti-VIP antibodies, the TH-positive and VIP-positive cell-bodies were also confirmed, as reported previously ( Fig. 1A; Hioki et al., 2010;Dougalis et al., 2012;Poulin et al., 2014). TH-positive cell bodies were located in the DR and linear nucleus regions. Most VIP-positive cell bodies were located in the ventral part of the periaqueductal region. Next, we examined the colocalization of DAT, TH, and VIP immunostaining in this region. The majority of the DAT DR-PAG neurons colocalized with TH or VIP signals (Fig. 1A,B; DAT1/TH1 59.9%, DAT1/VIP1 16.6%, and DAT1/TH1/VIP1 3.1% of the all DAT DR-PAG neurons; from four animals), only a few populations of DAT DR-PAG neurons were labeled with both TH and VIP. This result clearly indicates that the DAT DR-PAG neurons are heterogeneous. We examined the co-labeling of these three signals (DAT, TH, and VIP) for TH1 cells and for VIP1 cells in this region as well. Among the TH1 cells, 68.0% were positive for the DAT signal and 3.9% were positive for the VIP signal ( Fig. 1B; from four animals). Among the VIP1 cells, 61.0% were positive for the DAT signal and 12.4% were positive for the TH signal  1B; from four animals). It is commonly expected that DAT neurons also express TH, however, there is a significant portion of nonoverlapping of DAT-positive and TH-positive neurons. A possibility then arises that some off-target expressions of Cre-recombinase were introduced in VIP1/TH-neurons in DAT-Cre mice. To exclude this possibility, we also examined the colocalization of VIP immunoreactivity and DAT mRNA. Among the VIP1 neurons in the periaqueductal area, 53.6% were co-labeled with DAT in situ hybridization signal (Fig. 2 from two animals). Given that the sensitivity of in situ hybridization is regarded to be slightly lower than that of the Cre-loxP system, this result is likely to rule out the possibility of crerecombinase off-target expression and support the idea that there is a certain population of DAT DR-PAG neurons that express VIP but not TH.
The main target of DAT DR-PAG neurons is CeL and BNST OV To study the innervation pattern of DAT DR-PAG neurons, we specifically labeled DAT DR-PAG neurons with EYFP by injecting AAV5-EF1a-DIO-hChR2-EYFP into the DR-PAG region of DAT-Cre mice. Axonal fibers labeled with anti- GFP antibodies were located densely in the BNST OV and CeL. In contrast, fibers labeled were rarely observed in the striatum and basolateral amygdala (BLA), despite that these regions are known to receive dense DA innervation (Fig. 3A). The fiber signal intensities for the BNST OV , CeL, striatum, and BLA were 57.5 6 4.5, 74.5 6 9.9, 9.1 6 2.1, and 6.0 6 2.2 ( Fig. 3B; arbitrary units; quantified from four animals), respectively. The fiber densities in the BNST OV and CeL were significantly higher than those in the other regions ( Fig. 3B; permutation t test followed by Holm's correction; BNST OV versus striatum and BLA were statistically significant, p , 0.001 a and 0.001 b , respectively, nominal thresholds = 0.0083 and 0.01, respectively; CeL versus striatum and BLA were statistically significant, p , 0.001 c and 0.001 d , respectively; nominal thresholds = 0.0125 and 0.0167, respectively; for the other combinations p . 0.05 e,f ). We also tested the impact of DAT DR-PAG neuron ablations on the DAT signals in the CeL and BNST OV . For this purpose, we injected 6-OHDA into the DR-PAG region. 6-OHDA is known to ablate DA neurons (Breese and Traylor, 1972;Kirik et al., 1998;Blum et al., 2001). The intensities of DAT1 fiber fluorescence in the CeL and BNST OV were remarkably decreased by 6-OHDA injection, compared with the naive control case (Fig. 3C,D; normalized fluorescence intensity, arbitrary units; CeL, control 2.17 6 0.08 vs 6-OHDA 0.73 6 0.07, permutation t test, p , 0.001 g ; BNST OV , control 1.88 6 0.17 vs 6-OHDA 0.56 6 0.08, permutation t test, p , 0.001 h ; three animals for each group). These results indicate that the DAT DR-PAG neurons are the primary sources of DAT innervation in the CeL and BNST OV .

Properties of DR-PAG neurons that have projections to CeA or BNST
To retrogradely label the cell body of DR-PAG neurons projecting onto the extended amygdala, we made restricted small injections of FG centered in CeL or BNST OV . We investigated the colocalization of signals of FG, DAT (depending on tdTomato expression), and TH or VIP in DAT-Cre::Ai14 mice. The summary of injection sites for the CeA and BNST is shown in Figure 4. First, we examined the colocalization of FG, DAT, and TH, by injecting FG into the CeA or BNST. After FG injection into the CeA, 34.6% of the CeA projecting neurons were DAT-positive and 37.7% of the neurons were TH-positive ( Fig. 5A; from three animals). Additionally, 19.6% of the retrogradely labeled neurons were positive for both DAT and TH (FG/DAT/TH; Fig. 5A; from three animals). Similar to this, after FG injection into the BNST, 31.1% of the BNST projecting neurons were DAT-positive and 31.4% of the neurons were TH-positive ( Fig. 5B; from three animals). Additionally, 11.2% of the retrogradely labeled neurons were positive for both DAT and TH (FG/DAT/TH; Fig. 5B; from three animals). Next, we examined the colocalization of FG, DAT, and VIP. After FG injection into the CeA, 30.0% of the CeA projecting neurons were DAT-positive and 26.7% of the neurons were VIP-positive ( Fig. 6A; from three animals). Additionally, 16.1% of retrogradely labeled neurons were positive for both DAT and TH (FG/DAT/TH; Fig. 6A; from three animals). Among the VIP1 neurons innervating the CeA, 60.3% were DAT1 neurons. Similar to these results, after FG injection into the BNST, 31.4% of the BNST projecting neurons were DAT-positive and 15.5% of the neurons were VIP-  . DAT DR-PAG neurons innervate exclusively onto the CeL and BNST OV , and a vast majority of DAT fibers in these regions originated from the DR-PAG. A, Representative images of AAV5-EF1a-DIO-hChR2-EYFP injection onto the DR-PAG region of a DAT-Cre mouse. EYFP signal was visualized with anti-GFP antibody following DAB staining. Top left, Injection site of the AAV onto a DAT-Cre mouse. Many DAT1 neurons were observed in this region. Top middle left, Fiber terminals were restricted in the CeL. Top middle right, Fiber terminals were restricted in the BNST OV . Top right, No fiber was observed in the striatum. B, Paired mean differences for six comparisons are shown in the Cumming estimation plot. Signal intensities from four DAT-Cre mice are plotted on the upper axes; each paired set of observations is connected by a line. On the lower axes, each paired mean difference is plotted as a bootstrap sampling distribution. Mean differences are depicted as dots; 95% CIs are indicated by the ends of the vertical error bars. The fiber signal intensities in the BNST OV and CeL were significantly higher than those in other regions. C, Comparison of fiber fluorescence intensities in the CeL and BNST OV between control and 6-OHDA-injected mice. Control, In the CeL and BNST OV , dense DAT1 fibers were observed. 6-OHDA, In contrast, DAT1 fibers were barely observed. D, Mean differences for two comparisons are shown in the Cumming estimation plot. Signal intensities from three DAT-Cre::Ai14 mice are plotted on the upper axes; each mean difference is plotted on the lower axes as a bootstrap sampling distribution. Mean differences are depicted as dots; 95% CIs are indicated by the ends of the vertical error bars. There were statistically significant differences between control and 6-OHDA-injected mice. positive ( Fig. 6B; from three animals). Additionally, 10.5% of retrogradely labeled neurons were positive for both DAT and VIP (FG/DAT/VIP; Fig. 6B; from three animals). Among the VIP1 neurons innervating the BNST, 67.7% were DAT1 neurons. We also tested the colocalization of immunolabeled signals of FG, TH, and VIP. After FG injection into the CeA, 40.1% of retrogradely labeled neurons were positive for TH, 19.8% of the neurons were positive for VIP, and 1.4% of the neurons were positive for both TH and VIP (FG/TH/VIP; Fig. 7A; from four animals). After FG injection into the BNST, 40.2% of retrogradely labeled neurons were positive for TH, 17.3% of the neurons were positive for VIP, and 1.1% of the neurons were positive for both TH and VIP (FG/TH/VIP; Fig. 7B; from four animals). Altogether, the two subpopulations of DAT DR-PAG neurons, which are DAT/TH and DAT/VIP, had similar projection patterns onto the CeA and BNST.
The majority of VIP DR-PAG neurons that have projections onto the CeA or BNST are glutamatergic It has been suggested that the DA DR-PAG neurons corelease glutamate as well as dopamine (Matthews et al., 2016;Groessl et al., 2018). Since we have found that DAT DR-PAG neurons consisted of two subpopulations, which are TH-/VIP1 and TH1/VIP-, here comes a question whether both subpopulations are glutamatergic. To address this question, we examined the colocalization of  (Fig. 8B; three animals). These results suggested that the majority of VIP DR-PAG neurons projecting to the BNST/CeA are glutamatergic. We also tried to examine whether TH neurons in the DR-PAG region are VGlut2-positive or not in VGlut2-cre::Ai14 mice. However, the expression of tdTomato in the DR-PAG region, except the periaqueductal area where the majority of VIP neurons were observed, was not strong enough to quantify the colocalization of TH and VGlut2. At least, we confirmed that the majority of VIP DR-PAG neurons could be glutamatergic, innervating the CeA and BNST similarly. From the results shown in Figures 1, 2, and 6, these VIP DR-PAG neurons are likely to be DAT1/VIP1 neurons. Our experiment did not exclude the possibility that some TH DR-PAG neurons were also glutamatergic.
The extent of DAT expression in DAT DR-PAG neurons is relatively lower than that of DAT neurons in the ventral tegmental area (VTA) and the substantia nigra compacta (SNc) As already shown, many DAT-positive neurons exist in the DR-PAG region. This was confirmed in DAT-Cre::Ai14 mice (Figs. 1,5,6,9) and by RNA in situ hybridization (Figs. 2, 9). However, the number of DAT-antibody- positive cell-bodies in DR-PAG was obviously lower than that of the DAT-mRNA-positive cell-bodies or tdTomato-positive cell-bodies in DAT-Cre::Ai14 mice ( Fig. 9; DR-PAG). In contrast, a large number of DAT-antibody-positive cell-bodies were observed in the VTA and SNc regions, which was similar to that confirmed in DAT in situ hybridization and in DAT-cre:: Ai14 mice ( Fig. 9; VTA-SNc). Furthermore, the DAT-mRNA signal in the DR-PAG was less bold than that in the VTA-SNc.
These results suggest that the DAT protein translation is much lower in DAT DR-PAG neurons than in DAT VTA-SNc neurons, which may reflect the different roles of these DAT neuron populations.

Discussion
In this study, we confirmed that DAT DR-PAG neurons consist of the following two subpopulations: DAT1/TH1/ VIP-and DAT1/TH-/VIP1 in male mice. These two populations similarly innervate the CeL and BNST OV , which are the main targets of DAT DR-PAG neurons. Furthermore, the majority of VIP DR-PAG neurons projecting onto the CeA and BNST are suggested to be glutamatergic. This is the first study that identified two different subpopulations of DAT neurons in the DR-PAG region. These results implied possible functional heterogeneity among DAT DR-PAG neurons.

DAT DR-PAG neurons consist of two subpopulations
Consistent with previous reports (Poulin et al., 2018;Cardozo Pinto et al., 2019;Lin et al., 2020), we also confirmed that many DAT1 neurons exist in the DR-PAG region using DAT-Cre::Ai14 mice, DAT in situ hybridization, and anti-DAT-antibody staining. However, these DAT neurons were different from typical DAT neurons, such as DAT neurons in the VTA or SNc, since the protein expression of DAT at soma was much lower as shown in Figure 9. This result suggests that DAT DR-PAG neurons may have different functional characteristics from other DAT-positive neurons.
Since DAT is regarded as a reliable marker for DA neurons (Augood et al., 1993;Ciliax et al., 1999;Lewis et al., 2001), we first expected that basically all tdTomato1 neurons in DAT-Cre::Ai14 mice were TH-positive dopaminergic neurons. Indeed, more than half of tdTomato1 DAT DR-PAG neurons were TH-positive. However, a substantial amount of the DAT DR-PAG neurons, especially those located in the periaqueductal region, were TH-negative and VIP-positive. Previous studies focusing on mRNA expression have reported that all DAT, TH, and VIP mRNA were expressed in a single DAT DR-PAG neuron (Poulin et al., 2014(Poulin et al., , 2018La Manno et al., 2016;Hook et al., 2018;Tiklová et al., 2019). Also, TH and VIP signals were colocalized in the DR-PAG region in a study using TH-GFP mice (Dougalis et al., 2012). Our present findings, which are incompatible with these previous studies, would arise from the difference between mRNA and protein expression levels in these DAT DR-PAG neurons. mRNA translation is subject to multiple modulatory factors, and mRNA existence does not ensure a high level of protein expression. Consistent with our results, a study showed that the tdTomato1 neurons located in the periaqueductal region in DAT-Cre::Ai14 were TH-negative (Cardozo Pinto et al., 2019;Kramer et al., 2021). We ruled out the off-target expression of Cre-recombinase in DAT-Cre mice, showing that these TH-negative neurons were VIP neurons expressing DAT mRNA. Therefore, it was concluded that DAT DR-PAG neurons consist of two different subpopulations, which are DAT1/TH1/VIP-and DAT1/ TH-/VIP1.
As we have demonstrated, DAT DR-PAG neurons expressed lower amounts of DAT than other typical DAT neurons. Hence, the DAT1/TH1 subpopulation of DAT DR-PAG neurons could be DA neurons with a feature different from that of other DA neurons. Subpopulations of DA neurons in the VTA, which innervate the prefrontal cortex, BLA, and a part of the nucleus accumbens, express a lower amount of DAT than the DA neurons innervating the striatum (Lammel et al., 2008). Given that the DR-PAG TH1 neurons are regarded as a subgroup of the A10 cluster (Trulson et al., 1985;Descarries et al., 1986), DAT1/TH1 neurons in the DR-PAG might have similar characteristics to the mesocorticolimbic DA neurons described in the previous report (Lammel et al., 2008). Additionally, TH-positive DA neurons in the hypothalamus (including A11) barely express DAT (Barraud et al., 2010;Koblinger et al., 2014;Zhang et al., 2021). These DA neurons with low DAT expression should comprise a distinct subpopulation assigned to different functions than the canonical DA neurons. We also confirmed the presence of the DAT1/TH-/VIP1 group in the DR-PAG region. Since TH expression seems to be lower in this group, these DAT1/TH-/VIP1 neurons will not act as DA neurons. In addition, we have found this population is VGlut2-positive. Thus, DAT1/TH-/VIP1 neurons will use glutamate rather than DA as the primary neurotransmitter. DAT1/ TH-neurons have been also reported in various brain regions. In the A11 cluster, DAT1 fibers lacking TH were confirmed, even those were a tiny population (Koblinger et al., 2014). DAT1 neurons, lacking TH, in the ventral premammillary nucleus of the hypothalamus were glutamatergic (Soden et al., 2016). The noncanonical DAT neurons described in the present study might share similar functions with these DAT1/TH-neurons.
The projection of DAT DR-PAG neurons onto the extended amygdala Using FG retrograde labeling, we confirmed that DAT DR-PAG neurons innervate the CeA and BNST. Anterograde tracings and 6-OHDA lesion study revealed that DAT DR-PAG neurons exclusively innervate the extended amygdala, especially the CeL and BNST OV . Consistent with these results, recent studies taking advantage of Cre-driver lines demonstrated that TH or DAT neurons in the DR-PAG region have projections onto the CeA and BNST (Matthews et al., 2016;Groessl et al., 2018;Poulin et al., 2018;Lin et al., 2020). In addition, using FG retrograde labeling similar to the present study, it was reported that many TH neurons in the DR-PAG region have projections onto the extended amygdala (Hasue and Shammah-Lagnado, 2002). In the study by Hasue and Shammah-Lagnado, many retrogradely labeled TH neurons were observed not only in the DR-PAG region but also in other TH clusters such as A9 or VTA, while in our current study, the vast majority of TH-positive retrogradely labeled neurons were restricted in the DR-PAG region. We assume that this difference stemmed from the difference in the amount of FG injected. Our injection was small and well restricted in the CeA and BNST compared with the study by Hasue and Shammah-Lagnado (2002). Also, our results suggested that the DAT1/TH-/VIP1 subpopulation would have glutamatergic projections onto the CeA and BNST. Consistent with this, VIP innervations from the DR-PAG region to the BNST, which was not specified as DAT1/VGlut21, have been reported in several studies (Petit et al., 1995;Kozicz et al., 1998). This VIP signal in the BNST exhibits sexual dimorphism in humans, finches, and tree shrews (Zhou et al., 1995;Chung et al., 2002;Goodson et al., 2006;Ni et al., 2022). Yu et al. (2021) reported that the manipulation of DA DR-PAG neuronal activities induced different behavioral modulations between male and female mice, which might be because of the sexual dimorphism of VIP fibers in BNST, originating from the VIP subpopulation of DA DR-PAG neurons.
Adding to these accumulating reports, the present study confirmed that the two subpopulations of DAT DR-PAG neurons, TH1/VIP-and TH-/VIP1, similarly innervate the CeA and BNST. According to the spatial distribution of retrogradely labeled neurons, both subpopulations of DAT DR-PAG neurons seemed to bilaterally innervate the CeA and BNST. Consistent with this idea, single neuron tracing experiments demonstrated the existence of DAT DR-PAG neurons innervating both the CeA and BNST (Lin et al., 2020). Thus, the possibility would arise that these innervations onto the CeA and BNST might have a broad impact on neural activities in both regions simultaneously. Additional experiments will be required to clarify this possibility in the future.