Delineation of the PBP and A8 subpopulations.
The A10 neurons in the nonhuman primate extend the entire rostrocaudal extent of the midbrain (approximately 6 mm), and include the following subnuclei: the rostral linear nucleus (RLi), the caudolinear nucleus (CLi), the intrafascicular nucleus, the ventral tegmental nucleus (VTA), the paranigral nucleus, and the parabrachial pigmented nucleus (PBP, Halliday and Tork, 1986; McRitchie et al., 1995). In human and in monkey, the PBP is by far the largest sub-nucleus of the A10 (Halliday and Tork, 1986; Olszewski and Baxter, 2014), and sweeps dorsolaterally over the entire A9 region, to merge caudally with the A8 group. We therefore grouped all individual VTA subnuclei medial to the PBP as the ‘midline VTA’ group and designated the PBP as a separate A10 subnucleus. PBP TH-positive neurons are frequently larger than other A10 neurons, and oriented horizontally with long proximal dendrites in the mediolateral plane.
CaBP immunoreactivity is a reliable marker used for the determination of the A10 and A8 DA subregions across species (Lavoie and Parent, 1991; Gaspar et al., 1993; McRitchie et al., 1996; Haber and Fudge, 1997), due to high expression in many A10 and A8 DA cells, and a noticeable absence in the SNc (A9) DA cells (Figure 1A, rostrocentral, 1D, caudal). Large fiber paths (i.e., the oculomotor bundle, III, and medial lemniscus, ml) also remain unstained and serve as additional landmarks for subregional boundaries. By comparing the distribution of CRF-positive terminals with adjacent CaBP-labeled sections at the light microscopic level, robust CRF innervation was found predominately at rostrocentrally over the PBP, and at caudal levels over the RRF/A8 region (Figure 1B-C [rostrocentral], Figure 1E-F [caudal]; Kelly and Fudge, 2018). There were few labeled CRF+ fibers extending into the CaBP-negative A9 region. As expected, CRF-positive axonal terminals were also found in the midline VTA nuclei, and were of a higher density at rostrocentral compared to caudal levels (Figure 1B; 1E). In both the PBP and A8, dense patches of CRF-labeled fibers consisted of very thin, highly varicose fibers, as well as some thicker, beaded 'fibers en passant'( PBP, Figure 1C; RRF /A8, Figure 1F).
For region of interest (ROI) localization in EM studies, adjacent sections were immunostained for CRF, CaBP and TH (a marker for DA in the ventral midbrain; Pearson et al., 1983) (Figure 2). This approach allowed us to sample from areas with known dense termination of CRF positive fibers specifically in the PBP or A8, confirmed by matching microscopic anatomical landmarks (blood vessels patterns, *) for accurate excision (boxed region) of PBP and A8 blocks for EM samples.
Characteristics of CRF-labeled axon terminals and TH+ dendrites
At the EM level, CRF immunoreactivity was predominantly found in axon shafts and terminals (Figure 3) and some dendrites (not shown). Depending on the plane of sectioning, CRF immunoreactive axon terminals were presented as small spherical elements (cross sections, >0.2 microns) or elongated (oblique or longitudinal sections) structures containing punctate immuno-peroxidase deposits (See Figure 3A). Within immunoreactive axon terminals, the presence of both small, clear synaptic vesicles (SV; containing neurotransmitters) as well as large dense core vesicles (DCVs) was seen. The most prominent CRF immunoreactive terminals were associated with CRF+ DCVs in the range of 80-100 nm in size. Large CRF+ terminals often contained at least one immunoreactive DCV (Figure 3B). Unlabeled DCVs were also found within CRF+ and CRF- terminals, consistent with the presence of other neuropeptides in afferent terminals (Maley, 1990; VanBockstaele et al., 1996) . CRF positive axon terminals made contacts (defined as either asymmetric [black arrows] or symmetric [white arrows] type terminal appositions; Peters et al., 1991) denoting excitatory and inhibitory synapses (respectively) . Both types of synaptic profiles were found on TH- positive (DA) and TH-negative (non-DA) dendrites (Figure 3C). CRF-labeled axon terminals were also frequently enveloped by an astrocytic process (asterisks in Figure 3B, 3D) as documented by others (VanBockstaele et al., 1996). CRF+ axons that were closely apposed to dendritic elements but lacked a synaptic profile, were not counted.
TH immunoreactivity was visualized using gold/silver enhancement (Figure 4A-4D) which results in small electron dense deposits spaced throughout the cellular element, mainly in dendritic structures. Gold labeling was titrated so as to not obstruct classification of synaptic contacts, permitting visualization of large post-synaptic densities (PSDs) that suggest asymmetric (excitatory, black arrow heads) synapses, versus the thin PSDs that make up symmetric (inhibitory, white arrow heads) contacts. Asymmetric contacts are discernable by an obvious protein accumulation on the side of the postsynaptic compartment (PSD) with a thickness of 25-50 nanometers (black arrowheads). In contrast, symmetric contacts (inhibitory) have a slight electron-dense thickening associated with the postsynaptic membrane and are classified by the abutment of pre- and postsynaptic membranes with an absence of PSD accumulation (white arrowheads). The majority of TH immunoreactivity was in medium to large dendritic structures, characterized by an electron-lucent cytoplasm containing gold particles, multiple mitochondria, and PSDs at points of excitatory contacts. Single TH-positive dendrites often displayed multiple synaptic contacts, both symmetric and asymmetric in nature. Non-TH immunoreactive dendrites were identified as having similar ultrastructural features, but without gold particles. Given the diffuse distribution of gold particles throughout the dendritic structure, small offshoots (such as spines and thin dendritic elements may have been left unlabeled and therefore underrepresented in our analysis).
CRF contacts onto non-DA cells predominate in both PBP and A8
We first quantified CRF contacts on each cell type in the PBP and A8 (Figure 5A). The majority of CRF-positive contacts were onto non-DA cells in PBP (89%; 917 contacts: 1031 total contacts) and A8 (86%; 1029 contacts: 1202 total contacts) (Two-way ANOVA, F(1,14)=46.06, p<0001; PBP, black bars (PBP), p=0.0012; white bars (A8), p=0.0018, Tukey’s multiple comparisons test). Thus, the relative frequency of CRF contacts on DA versus non-DA cells was similar in the PBP and A8 (within DA comparisons: p=0.9985 n.s.; within non-DA comparisons: p=0.9288 n.s., two-way ANOVA with Tukey’s multiple comparisons test).
The proportion of CRF inhibitory contacts is greater than excitatory contacts in PBP and A8
We next quantified the proportion of either asymmetric (excitatory) or symmetric (inhibitory) contacts among both cell types in each region (Figure 5B). In the PBP and A8, symmetric contacts comprised the majority of all contacts (62% in the PBP, black bars, p=0.0036, Tukey’s multiple comparisons test; 63% in the A8, A8, white bars, p=0.0007, Tukey’s multiple comparisons test). There was a significant interaction of synapse type across both regions (asymmetric vs symmetric, two-way ANOVA, F(1,14)=44.69, p<0.0001). The number of each synaptic type was similar across regions (PBP vs A8, two-way ANOVA, F(1,14)=0.012, p=0.9146 n.s.) indicating that the balance of inhibitory versus excitatory contacts is similar in PBP and A8.
Characterizing type of CRF contact onto DA and non-DA cells types in the PBP and A8
We next investigated the nature of CRF interactions in the PBP versus A8 subregions, by quantifying the proportion of CRF-positive axon contacts across both cell type (DA vs non-DA) as well as synapse type (asymmetric [+] vs symmetric [-]). In the PBP (Figure 5C), there were significantly more synaptic contacts overall onto non-DA cells compared to DA positive cells (significant effect across cell type [DA versus non-DA], two-way ANOVA, F(1,12)= 103.7, p<0.0001). With respect to synapse type, we found that 89% of asymmetric contacts (black bars; 143 contacts: 162 total contacts) were found on non-DA cells (p=0.0002, two way-ANOVA with Tukey’s multiple comparisons tests). Similarly, 85% of symmetric contacts (gray bars; 185 contacts: 217 total contacts) were also found on non-DA cells (p<0.0001, Tukey’s multiple comparisons test). The proportion of symmetric vs asymmetric contacts did not differ with respect to DA versus non-DA cell type in this region (two-way ANOVA, F(1,12)=4.114, p=0.06).
In the RRF/A8 (Figure 5D), we also found significantly more overall synaptic contacts onto non-DA positive cells compared to DA positive cells, similar to the PBP (p=0.0003, Tukey’s multiple comparisons test). For the asymmetric contacts (black bars), 85% were onto non-DA cells (158 total contacts: 186 total overall contacts; p=0.0002; Tukey’s multiple comparisons test). Among the symmetric contacts (gray bars), we found that 88% of those contacts were also onto non-DA cells (281 total contacts: 321 overall contacts; p<0.0001, Tukey’s multiple comparisons test). We next analyzed differences in the proportion of asymmetric versus symmetric contacts across post-synaptic cell types (DA vs non-DA). Non-DA cells had significantly greater proportions of symmetric (inhibitory) type synapses compared to excitatory type synapses. The proportion of asymmetric vs symmetric contacts in the DA cell population was low and not significantly different from one another (p=0.9625, Tukey’s multiple comparisons test).
Frequency of CRF-positive axon contacts are comparable across males and females
CRF is differentially regulated in males and females via several mechanisms (Bangasser and Valentino, 2012; Bangasser et al., 2013). In females, stress induces enhanced CRF-mediated activation of the HPA axis and has enhanced effects on post-synaptic receptor dynamics (reviewed in, Bangasser and Valentino, 2012). To compare potential sex differences in the frequency of CRF-positive terminal interactions onto DA or non-DA neurons in the PBP and A8, we compared CRF-positive synaptic contacts across cell type in males (blue) and females (red) (Figure 6A). Similar to pooled cases (Figure 5), we noted significantly more contacts made on non-DA cells in males, with 93% of all contacts on non-DA cells in the PBP (585 contacts: 631 total contacts) and 86% of all contacts on non-DA cells in A8 (462 contacts: 538 total contacts). This was confirmed following statistical analysis (PBP, solid blue bar, p=0.0084; A8, hatched blue bar, p=0.05; Tukey’s multiple comparisons) with a significant effect for cell type (two-way ANOVA, F(1,10)=51.93, p<0.0001). In females, 83% and 91% of all contacts were on non-DA cells in PBP (332 contacts: 400 total contacts) and A8 (387 contacts: 423 total contacts), respectively, however, these no differences failed to reach statistical significance (solid red bar vs hatched red bar). CRF contacts in males and females were not significantly different for each cellular population (DA, males vs females, PBP, p=0.9999; DA, males vs females, A8, p=0.9987; non-DA, males vs females, PBP, p=0.3574; non-DA, males vs females, A8, p=0.9998).
We next compared the proportion of synapse type (asymmetric and symmetric) across DA and non-DA neurons in males and females for each region. In PBP (Figure 6B), there were more CRF+ contacts made onto non-DA cells regardless of synapse type in males (solid blue vs hatched blue, 92% of all asymmetric contacts were onto non-DA cells [69 contacts: 75 total asymmetric contacts], p=0.0408, 89% of all symmetric contacts were onto non-DA cells [87 contacts: 97 total symmetric contacts], p=0.0134) compared to DA cells. In females, 85% of all asymmetric contacts were onto non-DA cells (solid red vs hatched red, non-DA asymmetric: p=0.0491;74 contacts: 87 total asymmetric contacts) and 81% of all symmetric contacts were onto non-DA cells (solid red vs hatched red, non-DA symmetric: p=0.0159; 98 contacts:120 total symmetric contacts). We did not find significant differences between sexes when comparing within synapse type (blue solid vs red solid; hatched blue vs hatched red, for all comparisons p>0.9999). Given the disproportionate number of contacts onto non-DA cells in PBP, we further compared whether there were more asymmetric vs symmetric contacts within this subpopulation of cells. Here again, we did not find significant differences (males, blue hatched bars, p=0.9146; females, red hatched bars, p=0.78).
Similarly, in RRF/A8 (Figure 6C), we found 81% and 84% of all asymmetric (84 contacts: 104 total contacts) and symmetric contacts (164 contacts: 195 total contacts), respectively, were on non-DA cells compared to DA cells in males. In across-cell type comparisons, males did not show a significant difference in the proportion of asymmetric contacts on DA vs non-DA cells (solid blue bar vs hatched blue bar, p=0.06). In contrast, significantly more symmetric contacts were made onto non-DA cells compared to DA cells in males (blue bar vs hatched blue, p=0.0002). Females showed a similar trend with 89% of all asymmetric contacts (74 contacts: 83 total contacts) and 93% of all symmetric contacts (118 contacts: 126 total contacts) on non-DA cells suggesting an overall prevalence for synaptic interaction on the non-DA population of cells. These differences were highly significant in the female cohort, as shown in across-cell type comparisons (female asymmetric synapses ,solid red bar vs hatched red bar, p=0.0172; female symmetric synapses, solid red bar vs hatched red bar, p=0.0002; Tukey’s multiple comparisons test). To further assess the disproportionate number of contacts in non-DA cells in A8, we further compared whether there were more asymmetric vs symmetric contacts within this subpopulation of cells. We found significantly more symmetric synapses across both sexes in non-DA cells (males, blue comparison bar below graph, p=0.0171; females, red comparison bar below graph, p=0.0132). We did not see significant differences in male vs female proportions in RRF/A8.
While both DA subregions (PBP and A8) had the highest proportion of contacts onto non-DA cells, the proportion of asymmetric vs symmetric contacts within each of these cell populations was similar in most comparisons. This was not the case when comparing the asymmetric vs symmetric proportions in the A8 non-DA cells. Here we found (in both pooled [Figure 5] and sex comparison data [Figure 6]) significantly more symmetric (inhibitory) contacts onto non-DA cells in A8. These differences could be explained by the differential ratios of DA vs non-DA neurons in these regions, and its influence on our probability to record non-DA interactions in PBP (see Discussion). It might also suggest a unique CRF-mediated modulation within the A8 region.
CRF contacts are not influenced by sex or hormonal status in young macaques
How CRF expression and stress impact the DA system may be developmentally regulated (Coco et al., 1992; Izzo et al., 2005; Rincon-Cortes and Grace, 2017). Adolescent monkeys are transitioning into sexual maturation at three to five years of age, with changes in physical parameters (vaginal epithelial changes, testicular volume) as well as hormonal shifts (Plant, 2015). To determine if pubertal status or stress effects was a factor in our cohort, we performed gonadal hormone and cortisol assays from serum specimens on seven animals (four of which are included in this study) at several timepoints (Table 3). In serum preparations, analyte concentrations were comparable between males and females with expected differences between gonadal hormones. Stress hormones were not significantly different from one another across animals (Figure 7). These findings suggest that overall developmental and stress parameters were comparable across animals.