zPeriaqueductal Grey/Dorsal Raphe Dopamine Neurons Contribute to Sex Differences in Pain-Related Behaviors

SUMMARY

Sex differences in pain severity, response, and pathological susceptibility are widely reported, but the neural mechanisms that contribute to these outcomes remain poorly understood. Here, we show that dopamine (DA) neurons in the ventrolateral periaqueductal gray/dorsal raphe (vlPAG/DR) differentially regulate pain-related behaviors in male and female mice through projections to the bed nucleus of the stria terminalis (BNST). We find that activation of vlPAG/DR^DA+ neurons or vlPAG/DR^DA+ terminals in the BNST reduces nociceptive sensitivity during naïve and inflammatory pain states in male mice, while activation of this pathway in female mice led to increased locomotion in the presence of salient stimuli. We additionally use slice physiology and genetic editing approaches to demonstrate that vlPAG/DR^DA+ projections to BNST drive sex-specific responses to pain through DA signaling, providing evidence for a novel ascending circuit for pain relief in males and contextual locomotor response in females.

eTOC Blurb

Yu et al. demonstrate that dopamine neurons in the periaqueductal grey/dorsal raphe target the bed nucleus of the stria terminalis to reduce pain sensitivity in male mice and increase locomotion in female mice. Dopamine signaling is required for the sex-specific expression of these adaptive behaviors.

Graphical Abstract

INTRODUCTION

Men and women exhibit distinct responses to pain and this variable use of coping strategies contributes to the risk and severity of pathological pain conditions (Bartley & Fillingim, 2013). The midbrain plays a well-established role in these pain responses, as structures like the periaqueductal gray (PAG) and dorsal raphe (DR) have optimal anatomical positioning to integrate relevant contextual information with fight-or-flight behaviors (Wang & Nakai, 1994; Behbehani, 1995). The ventrolateral column of the PAG (vlPAG) and the neighboring DR are known generators of potent analgesia and enhanced opioid-induced anti-nociception (Hosobuchi et al., 1977; Mayer & Liebeskind, 1974; Fardin et al., 1984; Morgan et al., 1991; Li et al., 1993; Cai et al., 2014; McDevitt et al., 2014; Tovote et al., 2016). Activation of these structures elicits active and passive coping responses to the environment, enabling the use of adaptive anti-nociceptive or locomotor strategies to minimize the impact of pain (Bandler & Shipley, 1994; Maier et al., 1995). Subpopulations of vlPAG/DR have been identified as segregated drivers of these functions, with glutamate (Glu) neurons producing pain relief and escape behaviors in opposition to γ-aminobutyric acid (GABA) neurons (McDevitt et al., 2014; Tovote et al., 2016; Samineni et al., 2017; He et al., 2019; Seo et al., 2019; Zhu et al., 2019; Vaaga et al., 2020). These cell types have also been implicated in the sex-dependent modulation of pain through opioid signaling (Loyd et al., 2007; Schoo et al., 2008; Doyle et al., 2017). Considering that pain disproportionately affects women across the lifespan (Vetvik & MacGregor, 2017) and effective pain management is impeded by our limited knowledge on mechanistic drivers of pain in female subjects (Mogil, 2012; Bartley & Fillingim, 2013; Melchior et al., 2016; Gupta et al., 2017; Sorge & Totsch, 2017; Mogil, 2020), more insight on how vlPAG/DR regulates the expression of pain in male and female subjects is needed.

Dopamine (DA) neurons in the vlPAG/DR have been implicated in the modulation of pain sensitivity (Flores et al., 2004; Li et al., 2016). vlPAG/DR^DA+ neurons co-release Glu and play a similar role in anti-nociception (Flores et al., 2004; Li et al., 2016; Tovote et al., 2016; Taylor et al., 2019). Work from our lab and others have shown that vlPAG/DR^DA+ neurons are disinhibited by μ and κ opioid receptor agonists (Flores et al., 2004; Li et al., 2016; Li & Kash, 2019; Lin et al., 2020) and activated by ethanol (Li et al., 2013) to reduce thermal and mechanical nociceptive sensitivity (Taylor et al., 2019). These neurons exhibit additional contributions to arousal, fear learning, memory, reward, and locomotion (Flores et al., 2006; Lu et al., 2006; Matthews et al., 2016; Cho et al., 2017; Groessl et al., 2018; Porter-Stransky et al., 2019; Vaaga et al., 2020; Lin et al., 2020), with an overarching effect on incentivized responses to environmental salience (Cho et al., 2017; Lin et al., 2020). Although it remains unclear how vlPAG/DR^DA+ regulation of these functions come together to influence behavior, there is evidence for functional specificity of vlPAG/DR neurons based on downstream target regions (Tovote et al., 2016; Groessl et al. 2018), suggesting that vlPAG/DR^DA+ may concurrently recruit circuits to elicit a diverse set of behaviors in response to pain.

Previous characterizations of vlPAG/DR^DA+ and pain were performed exclusively in male subjects, leaving the function of these neurons undetermined in females. To address this disparity, the present study examines the role of vlPAG/DR^DA+ neurons in regulating pain responses for male and female mice. We specifically focus on vlPAG/DR^DA+ projections to the bed nucleus of the stria terminalis (BNST), since the BNST is a critical region for allocating emotional value to sensory information and likely plays an important role in pain (Hagiwara et al., 2013; Ide et al., 2013). Here, we used a combination of technical approaches, including chemogenetics, optogenetics, slice physiology, and CRISPR/Cas9-mediated knockout, to assess how vlPAG/DR^DA+ neurons and vlPAG/DR^DA+ projections to BNST impact nociceptive responses and related incentive behaviors. We found that vlPAG/DR neurons drive adaptive anti-nociceptive and locomotor behaviors differently in male and female mice via DA transmission to the BNST, with the context of neuronal activation being a key determinant of behavioral outcomes. These findings establish a critical role for vlPAG/DR^DA+ neurons in the sex-specific expression of pain-related behaviors.

RESULTS

vlPAG/DR^DA+ Neurons Produce Sex-Specific Reductions in Pain Sensitivity

Activation of vlPAG/DR^DA+ neurons reduces pain in male mice (Li et al., 2016; Taylor et al., 2019), but whether these anti-nociceptive effects apply to their female counterparts has yet to be determined. To probe vlPAG/DR^DA+ regulation of pain-related behaviors in both sexes, we first injected an adeno-associated virus carrying Cre-inducible hM3Dq (AAV8-hSyn-DIO-hM3Dq-mCherry; an excitatory designer receptor exclusively activated by designer drugs [DREADD; Armbruster et al., 2007]) or a control virus (AAV8-hSyn-DIO-mCherry) in the vlPAG/DR of adult male and female tyrosine hydroxylase (TH)-Cre mice (Figure 1A). This DREADD approach enabled systemic administration of clozapine-N-oxide (CNO) to promote hM3Dq signaling in vlPAG/DR neurons that express TH, a rate-limiting enzyme for DA biosynthesis (Figure 1B). Chemogenetic activation of these putative vlPAG/DR^DA+ neurons was employed during the measurement of pain sensitivity (Figure 1B), where the Hargreaves (heat exposure to the hind paw [6 replicates; Hargreaves et al., 1988]) and Von Frey (nylon monofilament exposure to the hind paw [SUDO method; Bonin et al., 2014]) tests were used to determine thermal and mechanical nociceptive sensitivity respectively. CNO treatment in hM3Dq mice reduced sensitivity to thermal (Figure 1C) and mechanical (Figure 1E) nociception compared to mCherry controls for males, but not in females (Figures 1D and 1F). Importantly, CNO-treated mCherry mice exhibited pain thresholds that were analogous to their saline-treated counterparts, suggesting that back-metabolism of CNO is an unlikely contributor to the observed effects in hM3Dq mice (Gomez et al., 2017). Together, these data show that chemogenetic activation of vlPAG/DR^DA+ neurons contribute to pain sensitivity in a sex-dependent manner, with the anti-nociceptive effects of the dopaminergic subpopulation being limited to male mice.

Figure 1. Activation of vlPAG/DR^DA+ Produces Sex-Specific Reductions in Pain Sensitivity.

(A) Chemogenetic approach for vlPAG/DR^DA+ activation. (Top) Diagram of virus infusion. (Bottom) Representative image of hM3Dq (orange), TH (green), and colocalization (yellow) in vlPAG/DR. Scale bar, 100 μm.

(B) Timeline with schematic of pain sensitivity testing and CFA treatment.

(C-D) Thermal nociceptive sensitivity of (C) male (n = 7) and (D) female (n = 7) TH-Cre mice following saline or CNO injection.

(E-F) Mechanical nociceptive sensitivity of (E) male (n = 7) and (F) female (n = 7) TH-Cre mice following saline or CNO injection.

(G-H) Post-CFA thermal nociceptive sensitivity in (G) male (n = 7) and (H) female (n = 7) TH-Cre mice following saline or CNO injection.

(I-J) Post-CFA mechanical nociceptive sensitivity in (I) male (n = 7) and (J) female (n = 7) TH-Cre mice following saline or CNO injection.

Data are shown as mean ±SEM. *p < 0.05; **p < 0.01; ***p < 0.001, ****p < 0.0001.

We next investigated the possibility of vlPAG/DR^DA+ intervention in pathological pain. Measurements of nociceptive sensitivity were repeated in the same cohort of TH-Cre mice following hind paw exposure to Complete Freund’s Adjuvant (CFA), a model of persistent inflammation, to determine the impact of vlPAG/DR^DA+ activation on heightened pain sensitivity. Similar to the effects observed in the naïve condition, CNO treatment in hM3Dq mice attenuated CFA-induced hyperalgesia for thermal and mechanical nociception in males (Figures 1G and 1I) but not females (Figures 1H and 1J). vlPAG/DR^DA+ anti-nociception was primarily active in the CFA-injected paw, with CNO treatment in male hM3Dq mice rescuing pain thresholds for the ipsilateral paw and having no effect on the contralateral paw (Figures S1A–S1D). Engagement of these neurons did not alter tail flick or avoidance behaviors in either sex (Figures S1E–S1P), in keeping with previous studies that have established a specific role for vlPAG/DR^DA+ neurons in the supraspinal modulation of pain (Li et al., 2016; Taylor et al., 2019). Consistent with this idea, our results demonstrate that chemogenetic activation of TH neurons in the vlPAG/DR exclusively produces anti-nociception in male mice during naïve and persistent inflammatory pain states.

Although chemogenetic activation of vlPAG/DR^DA+ neurons in female subjects did not produce statistically significant changes in pain sensitivity, a variable range of mechanosensitivity thresholds trending towards anti-nociception was observed in the Von Frey test (Figure 1F). It is posited that high variations in pain threshold can reflect differences in hormonal status, as fluctuations in testosterone and estrogen levels have been shown to alter DA availability and pain sensitivity (Di Paolo, 1994; Bradshaw et al., 2000; Purves-Tyson et al., 2012; Sorge et al., 2011; 2015). Therefore, it is possible that the functional contributions of vlPAG/DR^DA+ neurons were overshadowed in female mice by hormonal influences in our initial testing of pain sensitivity. To address this, we tested a separate cohort of female TH-Cre mice by tracking the impact of vlPAG/DR^DA+ activation on nociceptive sensitivity during different phases of the estrous cycle (Figure S2A). CNO treatment in hM3Dq mice resulted in comparable thermal and mechanical nociceptive sensitivity to mCherry controls regardless of hormonal status, suggesting that vlPAG/DR^DA+ function is not altered by estrous cycle (Figure S2B–S2C). Chemogenetic activation of vlPAG/DR^DA+ did appear to produce a trend for higher mechanosensitivity thresholds in the Von Frey test, however, where the mean of variable nociceptive responses was suggestive of an anti-nociceptive effect (Figure S2C; trend for main effect of Virus: p = 0.0504). This outcome is reminiscent of the observations from our initial testing of pain sensitivity, indicating that vlPAG/DR^DA+ activation in female mice reliably results in variable anti-nociception and that this variability cannot be explained by phase changes in the estrous cycle.

To assess if these differences in DREADD-evoked behavior were due to differences in cell numbers, we next used immunohistochemistry (IHC) and in situ hybridization (ISH) to quantify levels of TH protein and Th mRNA in the vlPAG/DR respectively and found no difference in expression between male and female mice (Figures S3A–S3B, S3E, and S3H). Additionally, we quantified dopamine beta-hydroxylase (DBH) as a marker for NE using DBH-eGFP reporter mice for IHC experiments or Dbh mRNA probes in C57BL/6J mice for ISH experiments. Contrary to observations in the vlPAG/DR of rats (Suckow et al., 2013), we observed no discernible overlap in TH and DBH expression to suggest the presence of noradrenergic neurons in the vlPAG/DR of mice (Figures S3B–S3G). These results collectively establish a role for vlPAG/DR^DA+ neurons in the sex-specific reduction of pain.

vlPAG/DR^DA+-BNST Drives Discrete Pain-Related Behaviors in Male and Female Mice

We next tested whether vlPAG/DR^DA+ projections to the BNST can alter pain-related behaviors by selectively expressing channelrhodopsin-2 (AAV5-EF1α-DIO-ChR2-eYFP) or a control fluorophore (AAV5-EF1α-DIO-eYFP) in the vlPAG/DR of adult male and female TH-Cre mice. Optical fibers were bilaterally implanted over the BNST, enabling 473 nm blue light stimulation of vlPAG/DR^DA+ terminals in the BNST during behavioral testing (Figures 2A–2B). Photostimulation of vlPAG/DR^DA+-BNST at 20 Hz in ChR2 mice caused a sustained reduction in thermal nociceptive sensitivity compared to eYFP controls for males but not females (Figures 2C–2D and S4A–S4B). This anti-nociceptive effect attenuated CFA-induced hyperalgesia as well, with sensitivity changes primarily exhibited in the ipsilateral paws of males (Figures 2G–2H and S4C–S4H). Activation of vlPAG/DR^DA+-BNST resulted in similar sex differential reductions in mechanical nociceptive sensitivity (Figures 2E–2F and 2I-2J) but did not affect tail flick responses (Figure S4I–S4J). These results suggest that the BNST is an important downstream target of vlPAG/DR^DA+ neurons for pain modulation, since the sex-specific anti-nociceptive effects of vlPAG/DR^DA+ neurons were able to be reproduced through activation of vlPAG/DR^DA+ terminals in the BNST.

Figure 2. vlPAG/DR^DA+-BNST Drives Anti-Nociceptive Behaviors in Male, but not Female, Mice.

(A) Optogenetic approach for vlPAG/DR^DA+ terminal activation in the BNST. (Left) Diagram of virus infusion and optical fiber implantation. (Right) AAV-DIO-ChR2 (green) expression in vlPAG/DR and BNST. Scale bar, 200 μm for vlPAG/DR; 100 μm for BNST.

(B) Schematic of photostimulation during thermal and mechanical nociceptive sensitivity testing. Epochs of laser exposure (i.e. trials [T1–8] for Hargreaves, days for Von Frey) are indicated for each assay.

(C-D) Thermal nociceptive sensitivity averaged by laser status in (C) male (n = 6–7) and (D) female (n = 8–10) TH-Cre mice.

(E-F) Mechanical nociceptive sensitivity averaged by laser status in (E) male (n = 7–8) and (F) female (n = 8–9) TH-Cre mice.

(G-H) Post-CFA thermal nociceptive sensitivity averaged by laser status in (G) male (n = 6–7) and (H) female (n = 4–6) TH-Cre mice.

(I-J) Post-CFA mechanical nociceptive sensitivity averaged by laser status in (I) male (n = 6–7) and (J) female (n = 4–6) TH-Cre mice.

Data are shown as mean ±SEM. *p < 0.05; **p < 0.01; ***p < 0.001, ****p < 0.0001.

In addition to the regulation of nociceptive sensitivity, vlPAG/DR^DA+ neurons have been implicated in a number of other functions, including (but not limited to) locomotion (Flores et al., 2006; Groessl et al., 2018), motivation (Flores et al., 2006; Matthews et al., 2016; Groessl et al., 2018; Lin et al., 2020), and salience (Lu et al., 2006; Cho et al., 2017; Porter-Stransky et al., 2019). Here, we determined whether vlPAG/DR^DA+ projections to BNST can modify performance in a variety of contexts by testing the same cohort of eYFP/ChR2 mice in a battery of behavioral assays. First, we evaluated the contributions of vlPAG/DR^DA+-BNST to pain-related functional impairment by introducing acute visceral nociception (via treatment with 1% acetic acid) during nesting, an innate behavior in mice that is temporarily impaired by pain and attenuated by analgesic drugs (Figure 3A; Negus et al., 2015). By activating vlPAG/DR^DA+ terminals in BNST during this assay, we were able to measure the projection’s contributions to visceral nociceptive (i.e. writhing, as defined by abdominal constriction / hind limb extension), affective-motivational (i.e. attenuation of impaired nesting), and locomotor behaviors related to pain. Although photostimulation of vlPAG/DR^DA+-BNST did not attenuate impaired nesting behavior (Figures 3B–3C), marked increases in writhing behavior (Figures 3D–3G) and locomotor activity were observed for females but not males (Figures 3H–3K). Since writhing relies on bodily movement, however, it is possible that the outcome was confounded by increases in locomotion.

Figure 3. vlPAG/DR^DA+-BNST Drives Pain-Related Locomotor Behaviors in Female, but not Male, Mice.

(A) Schematic of photostimulation during visceral nociceptive sensitivity testing.

(B-C) Cleared nesting zones with ddH2O/Acetic Acid treatment following optogenetic activation of vlPAG/DR^DA+-BNST in (B) male (n = 5) and (C) female (n = 5–6) TH-Cre mice.

(D-E) Writhing by Light ON and OFF sessions in (D) male (n = 5) and (E) female (n = 5–6) subjects.

(F-G) Writhing averaged by laser status in (F) male (n = 5) and (G) female (n = 5–6) subjects.

(H-I) Distance traveled by Light ON and OFF sessions in (H) male (n = 5) and (I) female (n = 5–6) subjects.

(J-K) Distance traveled by laser status in (J) male (n = 5) and (K) female (n = 5–6) subjects.

Data are shown as mean ±SEM. *p < 0.05; **p < 0.01; ***p < 0.001, ****p < 0.0001.

To differentiate these effects on locomotion from visceral nociception, we performed a series of follow-up assays. Using real-time place preference/aversion (RTPP/RTPA) and sociability with conspecific mice of the same and opposite sex, we were able to measure the rewarding, aversive, and locomotor contributions of the vlPAG/DR^DA+-BNST pathway in a variety of contexts that lack nociceptive stimuli (Figure 4). Generally, photoactivation of vlPAG/DR^DA+-BNST did not produce side or social preferences for males or females in the RTPP (Figures 4A–4B), sociability test with same sex conspecifics (Figures 4D–4F), sociability test with opposite sex conspecifics (Figures 4I–4K), or post-isolation sociability test with same sex conspecifics (Figures 4N–4P). However, vlPAG/DR^DA+-BNST stimulation did increase locomotion for female subjects in a context-dependent manner, with prototypically salient stimuli like environmental novelty and opposite-sex conspecific mice leading to greater increases in movement compared to more neutral stimuli (Figures 4C, 4G–4H, 4L-4M, and 4Q–4R). These data suggest that activation of vlPAG/DR^DA+ neurons drive locomotor responses to environmental salience through the BNST of female mice. Collectively, our results reveal that vlPAG/DR^DA+ projections to the BNST function in a sex-dependent manner, promoting adaptive behavioral responses to environmental salience by driving anti-nociception in males and context-dependent locomotion in females.

Figure 4. vlPAG/DR^DA+-BNST Increases Context-Dependent Locomotor Behaviors in Female, but not Male, Mice.

(A-C) Real-time place preference/aversion with optogenetic activation of vlPAG/DR^DA+-BNST.

(A) Schematic of RTPP/RTPA with designated sides for 20 Hz light stimulation (purple) and no stimulation (grey). Comparisons by sex are shown for (B) duration in stimulus zone (n = 5–6) and (C) distance traveled (n = 5–6).

(D-H) Sociability test with a same sex conspecific mouse and novel object. (D) Schematic of sociability test. (E-F) Ratio of mouse and object exploration by laser status and (G-H) distance traveled in (E, G) male (n = 7–8) and (F, H) female (n = 5–6) TH-Cre mice.

(I-M) Sociability test with an opposite sex conspecific mouse and novel object. (I) Schematic of sociability test. (J-K) Ratio of mouse and object exploration by laser status and (L-M) distance traveled in (J, L) male (n = 7–8) and (K, M) female (n = 8) TH-Cre mice.

(N-R) Sociability test with a male conspecific mouse and novel object following social isolation.

(N) Schematic of post-isolation sociability test. (O-P) Ratio of mouse and object exploration by laser status and (Q-R) distance traveled for (O, Q) male (n = 4–5) and (P, R) female (n = 5–6) TH-Cre mice.

Data are shown as mean ±SEM. *p < 0.05; **p < 0.01; ***p < 0.001, ****p < 0.0001.

Sex Differences in vlPAG/DR^DA+-BNST Transmission and Connectivity

Given the distinct behavioral outcomes observed for vlPAG/DR^DA+-BNST activation in male and female mice, we next examined the pathway’s functional properties using slice physiology. As previously noted in Li et al. (2016), photoactivation of vlPAG/DR^DA+ terminals in the BNST resulted in optically-evoked excitatory postsynaptic currents (oEPSCs) that exhibit onset properties of monosynaptic connections and optically-evoked inhibitory postsynaptic currents (oIPSCs) that exhibit the properties of polysynaptic connections (Figures 5A–5B and S5A). Comparisons by sex revealed no difference in the excitation/inhibition (E/I) ratio (Figure 5C), o(E/I)PSCsonset latency (Figures 5D and 5H), and o(E/I)PSCs amplitude (Figures 5E and 5I). By contrast, measures of paired-pulse ratio (PPR) revealed that females exhibited lower PPR for oIPSCs than males (Figure 5J), with no difference in oEPSC PPR (Figure 5F), suggesting a higher initial release probability of inhibitory polysynaptic connections in the BNST of female mice. Overall, there was a lower percentage of light-responsive BNST neurons for oIPSCs in female subjects (Figure 5K), but no difference in oEPSC responsivity (Figure 5G), when compared to males. Female mice thus possess less local functional connectivity for polysynaptic inhibitory connections between vlPAG/DR^DA+ terminals and BNST neurons than their male counterparts.

Figure 5. Sex Differences in vlPAG/DR^DA+-BNST Transmission and Connectivity.

(A) Experimental schematic illustrating whole cell patch clamp in ACSF. Recordings were performed in the BNST following optogenetic activation of vlPAG/DR^DA+ terminals.

(B) Representative o(E/I)PSC traces from BNST neurons following activation of vlPAG/DR^DA+ terminals with a single 1 ms pulse of 473 nm light (blue). Onset latency (green) and amplitude (pink) are indicated to highlight the distinct properties of excitatory and inhibitory transmission between vlPAG/DR^DA+ and BNST.

(C) E/I ratio comparison in male (n = 16 cells) and female (n = 5 cells) TH-Cre mice.

(D) Onset latency comparison for oEPSCs in male (n = 20 cells) and female (n = 14 cells) subjects.

(E) Amplitude comparison for oEPSCs in male (n = 20 cells) and female (n = 14 cells) subjects.

(F) Paired-pulse ratio of oEPSCs in male (n = 19–20 cells at each ISI) and female (n = 12–14 cells at each ISI) subjects.

(G) Percentage of responsive and non-responsive cells for oEPSCs in male (n = 52 cells) and female (n = 53 cells) subjects.

(H) Onset latency comparison for oIPSCs in male (n = 18 cells) and female (n = 5 cells) subjects.

(I) Amplitude comparison for oIPSCs in male (n = 18 cells) and female (n = 5 cells) subjects.

(J) Paired-pulse ratio of oIPSCs in male (n = 12–15 cells at each ISI) and female (n = 3–4 cells at each ISI) subjects.

(K) Percentage of responsive and non-responsive cells for oIPSCs in male (n = 52 cells) and female (n = 53 cells) subjects.

Data are shown as mean ±SEM. *p < 0.05; **p < 0.01; ***p < 0.001, ****p < 0.0001.

While sex differences in oEPSCs were not observed, it is possible that the fast transmission dynamics of Glu and/or GABA masked more subtle differences in volume transmission by DA. To more directly measure DA signaling in vlPAG/DR^DA+-BNST, we first combined optogenetics with fast-scan cyclic voltammetry to measure DA release from ChR2-expressing vlPAG/DR^DA+ terminals in the BNST (Figure 6A). Changes in DA release were observed in a pulse- and frequency-dependent manner. Light stimulation at 20 Hz produced progressive increases of DA following the introduction of more pulses (starting at 1 pulse and ending at 40 pulses), while frequencies ranging from 2 Hz to 30 Hz resulted in parabolic increases of DA, with 10 Hz stimulation yielding the greatest amount of DA release (Figure 6B). Comparisons by sex, however, revealed no differences in DA release or dopamine receptor D2 (D2R) regulation of release, regardless of light stimulation parameters (Figures 6B–6C).

Figure 6. Sex Differences in vlPAG/DR^DA+-BNST Dopaminergic Transmission.

(A) Experimental schematic illustrating fast-scan cyclic voltammetry in ACSF. Recordings were performed in the BNST following optogenetic activation of vlPAG/DR^DA+ terminals.

(B-C) Peak DA current following photostimulation of vlPAG/DR^DA+ terminals in the BNST in male (n = 6–7 slices in ACSF; n = 6 slices in ACSF + sulpiride) and female (n = 8 slices in ACSF; n = 6 slices in ACSF + sulpiride) TH-Cre mice. Recordings with (B) varying light pulses were performed in ACSF, then (C) repeated with varying frequencies in ACSF + 2 μM sulpiride. Drug effects were quantified using the percentage change in peak DA current following D2R antagonism.

(D) Experimental schematic illustrating whole cell patch clamp in ACSF + 3 mM kynurenic acid + 25 μM picrotoxin. Recordings were performed in the BNST following optogenetic activation of vlPAG/DR^DA+ terminals.

(E) Representative traces of BNST neurons responding to optically-evoked DA transmission from vlPAG/DR^DA+ terminals. (Left) 20 Hz (5 ms width, 20 pulses) stimulation in ACSF + kynurenic acid (KA) / picrotoxin (P) produces depolarization, hyperpolarization, or a mix of the two responses. (Middle-Right) These responses are blocked by the addition of SCH-23390 or sulpiride to ACSF + KA/P. Scale bars, x = time (4 sec), y = membrane potential (2 mV).

(F-I) Change in membrane potential of BNST neurons following optically-evoked DA transmission from vlPAG/DR. Comparison of (F) depolarization and (G) hyperpolarization size between male (depolarization: n = 26 cells; hyperpolarization: n = 13 cells) and female (depolarization: 23 cells; hyperpolarization: 14 cells) TH-Cre mice. DA receptor antagonism of (H) depolarization (n = 14 cells) and (I) hyperpolarization (n = 6 cells) is shown with individual cells from male (blue) and female (pink) subjects.

(J) Percentage of BNST neurons showing depolarization, hyperpolarization, and no response following light-evoked DA transmission from vlPAG/DR^DA+ terminals in male (n = 53 cells) and female (n = 56 cells) subjects.

Data are shown as mean ±SEM. *p < 0.05; **p < 0.01; ***p < 0.001, ****p < 0.0001.

Next, we measured DA transmission from vlPAG/DR to the BNST by repeating our optically-evoked slice physiology experiments with bath application of 3 mM kynurenic acid and 25 μM picrotoxin. This allowed us to block Glu and GABA activity and isolate the potential effects of DA release on downstream BNST neurons (Figure 6D). Following light stimulation at 20 Hz, three distinct types of physiological responses were observed: depolarization, hyperpolarization, or a mix of the two responses (Figure 6E). In some cases, changes in membrane potential were sustained and able to drive repeated action potentials lasting on the order of minutes (Figure S5C). When responses to DA transmission were quantified, sex differences for depolarization, but not hyperpolarization, were observed, with greater excitatory shifts in membrane potential appearing in male mice (Figures 6F–6G). These membrane effects were gated by DA receptors, with depolarization being sensitive to dopamine receptor D1 (D1R) antagonism and hyperpolarization being sensitive to D2R antagonism (Figures 6H–6I). Despite sex-specific mRNA expression of Drd2 but not Drd1 in the BNST (Figure S5B), D1R antagonism was greater in male mice (Figure 6H), while D2R receptor antagonism did not differ in males and females (Figure 6I). Overall, the percentage of depolarizing, hyperpolarizing, and non-responsive cells did not vary by sex (Figure 6J). Treatment with CFA did not alter vlPAG/DR^DA+-BNST transmission or connectivity compared to saline-treated controls (Figures S6 and S7). Taken together, these results indicate that sexual dimorphism in DA transmission is a notable feature of the vlPAG/DR^DA+-BNST pathway, with male mice showing greater D1R-gated depolarization than females.

vlPAG/DR^DA+ Anti-Nociception and Locomotion is Dopamine-Dependent

To determine the necessity of DA for vlPAG/DR^DA+ function, we combined chemogenetics with CRISPR/Cas9-mediated genome editing by co-injecting AAV-hSyn-DIO-hM3Dq-mCherry with either a TH CRISPR (AAV1-DIO-saCas9-U6-sgTH) or control (AAV1-DIO-saCas9-U6-sghTH; CTRL) virus in the vlPAG/DR of adult male and female TH-Cre mice (Figure 7A). This approach enabled us to functionally characterize vlPAG/DR^DA+ neurons following conditional genetic knockout of the DA biosynthetic enzyme TH (Hunker et al., 2020). Validation of the TH CRISPR virus revealed a 36%−46% reduction of TH expression in the vlPAG/DR of male and female subjects (Figure 7A). Using the Hargreaves and Von Frey tests, we were able to reproduce the anti-nociceptive effects of vlPAG/DR^DA+ activation, where CNO treatment reduced thermal and mechanical nociceptive sensitivity in hM3Dq+CTRL mice for males but not females (Figures 7C–7F). This effect was attenuated in hM3Dq+TH CRISPR mice, as reductions in TH mitigated the anti-nociceptive effects in males without baseline sensitivity changes in either sex (Figures 7C–7F). These data indicate that the presence of TH in vlPAG/DR is necessary for the sex-specific anti-nociceptive effects of vlPAG/DR^DA+ activation.

Figure 7. vlPAG/DR^DA+ Anti-Nociception and Locomotion is Dopamine-Dependent.

(A) Chemogenetic approach for vlPAG/DR^DA+ activation with local genetic deletion of TH. (Top-Left) Diagram of virus infusion. (Top-Right) Detailed schematic of TH CRISPR provided by Zweifel lab. (Bottom-Left) Validation of TH deletion in vlPAG/DR for male (n = 7) and female (n = 7) TH-Cre mice. (Bottom-Right) hM3Dq expression (orange), TH immunoreactivity (green), and colocalization (yellow) in vlPAG/DR. Scale bar, 100 μm.

(B) Schematic illustrating the measurement of thermal and mechanical nociceptive sensitivity and locomotion following treatment with saline or CNO.

(C-D) Thermal nociceptive sensitivity of (C) male (n = 7) and (D) female (n = 7) TH-Cre mice following saline or CNO injection.

(E-F) Mechanical nociceptive sensitivity of (E) male (n = 7) and (F) female (n = 7) TH-Cre mice following CNO or saline injection.

(G-H) Locomotor activity of (G) male (n = 7) and (H) female (n = 7) TH-Cre mice following a saline injection.

(I-J) Locomotor activity of (I) male (n = 7) and (J) female (n = 7) TH-Cre mice following a CNO injection.

(K) Pre-Saline Habituation (I). Averaged locomotor activity prior to saline treatment in male (n = 7) and female (n = 7) TH-Cre mice.

(L) Saline (I). Averaged locomotor activity after saline treatment in male (n = 7) and female (n = 7) TH-Cre mice.

(M) Pre-Saline Habituation (II). Averaged locomotor activity prior to CNO treatment in male (n = 7) and female (n = 7) TH-Cre mice.

(N) CNO (II). Averaged locomotor activity after CNO treatment in male (n = 7) and female (n = 7) TH-Cre mice.

Data are shown as mean ±SEM. *p < 0.05; **p < 0.01; ***p < 0.001, ****p < 0.0001.

Considering our earlier findings that vlPAG/DR^DA+ activation in the presence of salient contextual features selectively drives locomotion in females, we next assessed the effect of TH knockout on locomotor activity in a novel environment. We tested the same cohort of hM3Dq+CTRL and hM3Dq+TH CRISPR mice in an open field, where subjects were free to explore the apparatus for an initial 30-minute habituation phase. Subjects were then treated with saline or CNO before being returned to the apparatus for an additional 90-minute test phase. TH knockout reduced locomotion in a sex-independent manner during the initial 30-minute habituation phase (Figures 7G–7H and 7K). However, no locomotor changes were observed in subsequent habituation (Figures 7I–7J and 7M) or test phases (Figures 7G–7J, 7L, and 7N) following treatment with saline or CNO. These data provide evidence that TH is necessary for vlPAG/DR neurons to promote novelty-induced locomotion during the initial exposure of an apparatus. Taken together, the ability of vlPAG/DR^DA+ neurons to meet environmental demands such as pain and contextual novelty is evidently dependent on the presence of TH, suggesting that DA originating in the vlPAG/DR is necessary for the sex-specific expression of ant-inociceptive and locomotor behaviors.

Finally, we tested whether TH knockout in vlPAG/DR can prevent the anti-nociceptive and locomotor effects of systemic morphine administration (3 mg/kg) in hM3Dq+CTRL and hM3Dq+TH CRISPR mice (Figure 8A), since vlPAG/DR^DA+ neurons are disinhibited by morphine and may contribute to variable drug effects in males and females. Contrary to this hypothesis, TH knockout did not affect morphine anti-nociception in males or females (Figures 8B–8E). Similar results were observed for locomotion, as sex-dependent changes in behavior were not evident for the habituation (Figures 8F–8G, 8J, and 8L) or test phases (Figures 8H–8I, 8K, and 8M) following treatment with saline or morphine. These results collectively suggest that intact DA signaling is necessary for the sex-specific anti-nociceptive and locomotor effects of vlPAG/DR activation, while systemic morphine recruits extra-vlPAG/DR^DA+ mechanisms for the expression of these adaptive behaviors.

Figure 8. Morphine Anti-Nociception and Locomotion is Not Dependent on vlPAG/DR Dopamine.

(A) Pharmacological approach for opiate-induced vlPAG/DR^DA+ activation with local genetic deletion of TH. Schematic illustrating measurement of thermal and mechanical nociceptive sensitivity and locomotion following systemic treatment with saline or morphine.

(B-C) Thermal nociceptive sensitivity of (B) male (n = 7) and (C) female (n = 7) TH-Cre mice following saline or morphine injection.

(D-E) Mechanical nociceptive sensitivity of (D) male (n = 7) and (E) female (n = 7) TH-Cre mice following saline or morphine injection.

(F-G) Locomotor activity of (F) male (n = 7) and (G) female (n = 7) TH-Cre mice following a saline injection.

(H-I) Locomotor activity of (H) male (n = 7) and (I) female (n = 7) TH-Cre mice following a morphine injection.

(J) Pre-Saline Habituation (III). Averaged locomotor activity prior to saline treatment in male (n = 7) and female (n = 7) TH-Cre mice.

(K) Saline (III). Averaged locomotor activity after saline treatment in male (n = 7) and female (n = 7) TH-Cre mice. female (n = 7) TH-Cre mice.

(L) Pre-Morphine Habituation (IV). Averaged locomotor activity prior to morphine treatment in male (n = 7) and female (n = 7) TH-Cre mice.

(M) Morphine (IV). Averaged locomotor activity after morphine treatment in male (n = 7) and female (n = 7) TH-Cre mice.

Data are shown as mean ±SEM. *p < 0.05; **p < 0.01; ***p < 0.001, ****p < 0.0001.

DISCUSSION

The vlPAG/DR has been implicated in a variety of pain-related behaviors, including anti-nociception, escape, and other functions necessary for survival (Hosobuchi et al., 1977; Mayer & Liebeskind, 1974; Fardin et al., 1984; Morgan et al., 1991; Li et al., 1993; Bandler & Shipley, 1994; Maier et al., 1995; Loyd & Murphy, 2006; Linnman et al., 2012; Cai et al., 2014; Seo et al., 2019; Wright & McDannald, 2019). Although it is widely accepted that the expression and experience of pain can vary by sex, little is known about the neural mechanisms that contribute to these behaviors (Aloisi et al., 2007; Sorge et al., 2011; Mogil, 2012; Posillico et al., 2015; Mapplebeck et al., 2016; Doyle et al., 2017; Dance et al., 2019; Inyang et al., 2019). In the present study, we define a midbrain-to-extended amygdala circuit responsible for the generation of divergent pain-related behaviors in male and female mice. Expanding on our findings from Li et al. (2016), we report that activation of vlPAG/DR^DA+ neurons or projections to BNST reduces thermal and mechanical nociceptive sensitivity in male mice. In accordance with Taylor et al. (2019), this effect persists after CFA treatment, with DA neurons exhibiting additional attenuation for sensitivity increases associated with long-term inflammation Surprisingly, vlPAG/DR^DA+-BNST activation did not result in anti-nociceptive effects for female mice, instead producing robust locomotor behaviors in a context-dependent manner. Reductions in local TH attenuated the sex-specific behavioral effects of vlPAG/DR^DA+ activation, possibly reflecting the distinct mechanisms of DA transmission and polysynaptic inhibition observed in the BNST of male and female mice. Collectively, these findings provide significant new insight into the functional role of vlPAG/DR^DA+ neurons and how projections to the BNST contribute to sex differences in pain.

In support of earlier evidence that vlPAG/DR^DA+ anti-nociception depends on both D1R and D2R signaling (Taylor et al., 2019), we found that DA transmission to the BNST generates a mix of depolarizing and hyperpolarizing responses that are gated by D1R and D2R respectively. More robust D1R-gated depolarization and D1R antagonism were observed in male mice, suggesting that D1R signaling contributes to sex differences in pain sensitivity (Flores et al., 2004; Megat et al., 2018). D1R antagonism in the BNST has previously been shown to exacerbate formalin-induced nociception in a sex-specific manner (Hagiwara et al., 2013). This earlier work, however, reported stronger effects in female rats, contradictory to the DA-dependent anti-nociception we observed in male mice. Considering that our results were limited to DA signaling from a specific input, it is possible that the direction of sex-specific anti-nociception depends on the source of DA (Glangetas & Georges, 2016). Determining the influence of phasic/escapable (e.g. heat, pressure) and tonic/inescapable (e.g. acetic acid, formalin) types of pain on DA inputs to the BNST may provide further clarification as well (Morgan & Franklin, 1990; Altier & Stewart, 1999; Taylor et al., 2016). Regardless of these distinctions, our conclusion that activating vlPAG/DR^DA+ projections to BNST reduces acute and persistent inflammatory pain provides further evidence that DA signaling can enhance anti-nociception to counteract severe pain (Burrill et al., 1944; Goetzl et al., 1944; Ivy et al., 1944; Dennis and Melzack, 1983; Chudler & Dong, 1995; Altier & Stewert, 1999; Magnusson & Fisher, 2000; Wood et al., 2007; Hagiwara et al., 2013; Megat et al., 2018).

Although activation of vlPAG/DR^DA+ neurons and its projections to BNST produced negligible effects on nociceptive sensitivity in female mice, the assessment of visceral nociception and valence-specific behaviors revealed an additional role for the projection in sex- and context-dependent locomotion. Consistent with previous reports that vlPAG/DR^DA+ increases locomotion in a neutral environment depending on the parameters of activation (Correia et al., 2017; Groessl et al., 2018; Taylor et al., 2019), we found that TH in vlPAG/DR is necessary to drive locomotion during the first habituation phase of an open field test for both males and females, despite showing no effects with CNO administration. In the presence of visceral nociception, however, optogenetic activation of vlPAG/DR^DA+-BNST selectively increases locomotion in females. A similar phenotype is exhibited in the sociability test, with pathway activation driving locomotor increases in the presence of opposite, but not same, sex mice. Considering that vlPAG/DR^DA+ activity is modulated by salience (Cho et al., 2017; Porter-Stransky et al., 2019; Cho et al., 2020), it is possible that this sex-specific locomotor phenotype is driven by the synergistic effects of pathway activation and context (Liu & Dan, 2019; Cazettes et al., 2020). Interestingly, the allocation of salience in a novel environment is thought to be modulated by DA signaling in the BNST, with D2R antagonism reducing discriminative learning for safety- and threat-related stimuli (De Bundel et al., 2016). Thus, it remains an intriguing possibility for D2R in the BNST to regulate salience, as the receptor may act as a mechanism to alter how context impacts vlPAG/DR^DA+ locomotion in males and females. Evidence for sex-specific modulation of anti-nociceptive (Hagiwara et al., 2013) and locomotor behaviors (Monleon et al., 1998; Schindler & Carmona, 2002) by DA receptor antagonism in the BNST supports this possibility. However, a limitation of this study is that in vivo characterization of DA-R antagonism in the vlPAG/DR^DA+-BNST pathway was not performed, so it remains unclear how blocking the downstream effects of D1R vs. D2R recruitment in the BNST impacts these sex-specific behaviors.

vlPAG/DR^DA+ neurons likely generate discrete pain responses through distinct third-order neurons, targeting BNST neurons with outputs to structures responsible for anti-nociception, locomotion, or a combination of these functions (e.g. central nucleus of the amygdala [CeA], ventral tegmental area, parabrachial nucleus; Dong et al. ,2001; Lebow & Chen, 2016). Shifts in the strength of these functionally distinct BNST circuits are anticipated as environmental demands differentially modulate the recruitment of vlPAG/DR^DA+ neurons in males and females. Although a molecular identity for vlPAG/DR^DA+-connected BNST neurons has yet to be determined, several markers in the BNST have exhibited roles in sex-specific responses to stressors and other environmental demands (Janitzky et al., 2014; Klampfl et al., 2016; Smithers et al., 2019) that may be regulated by local DA interactions (Kash et al., 2008; De Bundel et al., 2016). Of particular interest are protein kinase C delta (PKC-δ) and somatostatin (SOM) neurons, two segregated subpopulations in the BNST that receive input from nociceptive structures (Ye & Veinante, 2019), respond to peripheral inflammation (Wang et al., 2019), and diverge in colocalization with D2R to modulate salience (De Bundel et al., 2016). vlPAG/DR^DA+ inputs activate both PKC-δ and SOM neurons in the CeA (Groessl et al., 2018), which share reciprocal connections with PKC-δ and SOM neurons in the BNST (Ye & Veinante, 2019), suggesting that these downstream structures may exhibit mutual regulation through specific subsets of cells. Collectively, these sex differences in BNST outputs and cell type are promising candidates to explain the functional divergences observed in this pathway. Whether downstream BNST neurons similarly exhibit a specific composition of efferents or molecular identities to determine how vlPAG/DR^DA+ inputs generate sex-specific anti-nociceptive and locomotor behaviors requires further investigation.

The discovery of sex-specific mechanisms for pain responses has major implications for human health. Considering that pain disproportionately affects women, the identification of a novel mechanism for sex-specific responses to pain may explain why disparities in susceptibility exist and indicate a feasible path towards improving analgesia by sex. Having demonstrated that vlPAG/DR^DA+ neurons are inconsistent drivers of anti-nociception for female mice, with no detectable contributions of the estrous cycle, we posit that vlPAG/DR inputs to the BNST can promote analgesia in females. However, mechanistic differences in DA transmission and polysynaptic inhibitory connections may predispose these neurons to drive more robust locomotor behaviors. This interpretation of vlPAG/DR^DA+ function is complicated by the fact that increases in locomotion can falsely diminish reflexive response times, resulting in confounded readouts of nociceptive thresholds. It is thus possible that vlPAG/DR^DA+ neurons concurrently recruit anti-nociceptive and locomotor effects, as increased locomotion would neutralize any measurable changes in pain sensitivity that are happening at the same time. Since hM3Dq- or ChR2-expressing female mice show increased locomotion compared to controls, a logical assumption to make would be that vlPAG/DR^DA+ activation drives female mice to move more during pain testing, resulting in faster paw withdrawal times – or “greater sensitivity to pain” – compared to controls, a phenotype that we have failed to see in females despite repeated nociceptive testing. Therefore, it is unlikely that a lack of analgesic effect in females is confounded by locomotion, but more explicit efforts to parse out these functions are required. Additional considerations for these multifaceted mechanisms in vlPAG/DR^DA+ function may be of interest for the development of therapeutic interventions, as the ability to selectively tap into anti-nociception and locomotion would be advantageous for reducing the impact of pain in a variety of modalities and contexts.

Current treatment for severe pain requires opioids, but these drugs exhibit sex biases in their efficacy (Kest et al., 2000; Fillingim & Gear, 2004; Bernal et al., 2007; Loyd et al., 2007; Loyd et al., 2008; Craft, 2008; Posillico et al., 2015). Since vlPAG/DR^DA+ neurons are endogenously recruited by morphine (Flores et al., 2006; Li et al., 2016; Lin et al., 2020) and drive sex-specific responses to pain, vlPAG/DR^DA+ neurons may be a promising candidate for opioid modulation of pain-related behaviors. In the present study, we report that reducing TH expression by 36%−46% in the vlPAG/DR mitigates the anti-nociceptive and locomotor effects of vlPAG/DR^DA+ activation, but not systemic morphine, in male and female mice. Since morphine also disinhibits Glu neurons in the vlPAG/DR to produce anti-nociceptive and freezing behaviors, it is possible that these extra-DA mechanisms of μ opioid receptor activation overshadowed the impact of our modest TH knockout in the vlPAG/DR. Previous studies have shown that an approximate 62% loss of local DA neurons was required to attenuate morphine anti-nociception (Flores et al., 2004), while an estimated 50% reduction in TH was needed to block morphine-induced conditioned place preference/aversion (Lin et al., 2020). These data indicate that graded amounts of DA impairment in vlPAG/DR^DA+ neurons may be necessary to alter the expression of nociceptive and affective-motivational behaviors associated with opioid exposure, since the TH deletion we observed was only able to prevent the sex-specific effects of vlPAG/DR^DA+ activation and not the broader effects of morphine. This distinction suggests that selective targeting of vlPAG/DR^DA+ activation could be a more reliable means of impacting the effects of opioid exposure than systemic approaches.

With these findings in mind, it remains a possibility that vlPAG/DR^DA+ projections to BNST contribute to opioid modulation of pain responses. Supporting this notion, morphine increases DA in the BNST (Carboni et al., 2000), with converging DA and Glu release activating pERK via D1R (Valjent et al., 2005). Consistent with these data, we found that activation of vlPAG/DR inputs to BNST produces anti-nociceptive effects in males but not females, a sex-specific effect that may be explained by D1-gated depolarization in the BNST. Notably, the analgesic effects of vlPAG/DR^DA+ were achieved without promoting reward/aversion (McDevitt et al., 2014). Systemic increases in DA have been shown to reduce the amount of opioids needed for anti-nociception while reducing side effects like sedation, loss of alertness, and cognitive deficits in humans (Forrest et al., 1977; Bruera et al. 1992; Dalal & Melzack, 1998), so vlPAG/DR^DA+ projections to BNST may be a unique contributor to this effect of DA signaling and opioid anti-nociception in males. Considering that women experience μopioid analgesia at the expense of more extreme side effects (Dahan et al., 1998; Sarton et al., 2000; Zacny , 2001; Cepeda & Carr, 2003; Miller & Ernst, 2004), we believe that learning more about the vlPAG/DR^DA+-BNST pathway and parallel circuits in the midbrain will help generate novel approaches to achieving comparable analgesia in females and minimize the access gap for effective pain management.

STAR METHODS

RESOURCE AVAILABILITY

LEAD CONTACT

Further information and request for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Thomas L. Kash (thomas_kash@med.unc.edu).

MATERIALS AVAILABILITY

This study did not generate new unique reagents.

DATA AND CODE AVAILABILITY

The published article includes all data generated or analyzed during this study.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

All animals were bred inhouse and maintained on a 12-h light/dark cycle (light on at 7:00, light off at 19:00) with rodent chow and water available ad libitum. Male and female TH-Cre mice (aged 6–12 weeks, C57BL/6J background) were group-housed with same-sex littermates until surgery or CFA treatment. All procedures were performed in accordance with the NIH Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee at UNC Chapel Hill.

TH-Cre Details

The mouse strain used for this research project, STOCK Tg(Th-cre)FI172Gsat/Mmucd, RRID:MMRRC_029177-UCD, was obtained from the Mutant Mouse Resource and Research Center (MMRRC) at University of California at Davis, an NIH-funded strain repository, and was donated to the MMRRC by Nathaniel Heintz, Ph.D., The Rockefeller University, GENSAT and Charles Gerfen, Ph.D., National Institutes of Health, National Institute of Mental Health.

Number of Animals per Experiment

-Figure 1 and Supplemental Figure 1: TH-Cre mice (Male mCherry control [n = 7], Male hM3Dq [n = 7], Female mCherry control [n = 7], Female hM3Dq [n = 7]).

-Supplemental Figure 2: TH-Cre mice (Female mCherry control in Estrus [n = 7], Female mCherry control in Non-Estrus [n = 7], Female hM3Dq in Estrus [n = 7], Female hM3Dq in Non-Estrus [n = 7]).

-Supplemental Figure 3: IHC Cohort #1 (S3A) used C57BL/6J mice from TH-Cre breeders (Males [n = 7], Females [n = 7]); IHC Cohort #2 (S3B-S3D) used DBH-eGFP mice (Males [n = 4], Females [n = 4]); ISH Cohort (S3E-S3G) used DBH-eGFP mice (Males [n = 4], Females [n = 2]).

-Figure 2 and Supplemental Figure 4: TH-Cre mice (Male eYFP [n = 7], Male ChR2 [n = 6–8], Female eYFP [n = 9–10], Female ChR2 [n = 8]).

-Figure 3: TH-Cre mice (Male eYFP [n = 5], Male ChR2 [n = 5], Female eYFP [n = 5], Female ChR2 [n = 6]).

-Figure 4: RTPP/RTPA (4A-4C) used TH-Cre mice (Male eYFP [n = 4], Male ChR2 [n = 5], Female eYFP [n = 5], Female ChR2 [n = 6]); Sociability test with same sex conspecifics (4D-4H) used TH-Cre mice (Male eYFP [n = 7], Male ChR2 [n = 8], Female eYFP [n = 5], Female ChR2 [n = 6]); Sociability test with opposite sex conspecifics (4I-4M) used TH-Cre mice (Male eYFP [n = 7], Male ChR2 [n = 8], Female eYFP [n = 8], Female ChR2 [n = 8]); Post-isolation sociability test with same sex conspecifics (4N-4R) used TH-Cre mice (Male eYFP [n = 4], Male ChR2 [n = 5], Female eYFP [n = 5], Female ChR2 [n = 6]).

-Figure 5 and Supplemental Figure 5: TH-Cre mice used for E/I transmission experiments (Male ChR2 [n = 5–6], Female ChR2 [n = 3–4]).

-Figure 6 and Supplemental Figure 5: TH-Cre mice used for DA transmission experiments (Male ChR2 [n = 4], Female ChR2 [n = 5–6]).

-Supplemental Figures 6 and 7: TH-Cre mice used for E/I transmission experiments (Male ChR2 [n = 4–5], Female ChR2 [n = 5]). TH-Cre mice used for DA transmission experiments (Male ChR2 [n = 2–4], Female ChR2 [n = 3–4]).

-Figures 7 and 8: TH-Cre mice (Male hM3Dq + sghTH control [n = 7], Male hM3Dq + TH CRISPR [n = 7], Female hM3Dq + sghTH control [n = 7], Female hM3Dq + TH CRISPR [n = 7]).

METHOD DETAILS

Stereotaxic Surgery

Adult mice (> 6 weeks of age) were anesthetized with isoflurane (1–3%) in oxygen (1–2 l/min) and aligned on a stereotaxic frame (Kopf Instruments, Tujunga, CA). All surgeries were conducted using aseptic techniques in a sterile environment. Microinjections were performed with a 1 μl Neuros Hamilton syringe (Hamilton, Reno, NV) and a micro-infusion pump (KD Scientific, Holliston, MA) that infused virus at 100 nl/min. Viruses were administered unilaterally at an angle of 20° in the middle-posterior region of vlPAG/DR (450 nl for all experiments [exception: 500 nl for Figure 2 experiments]; relative to bregma: ML 0.00 mm, AP −4.50 mm, DV −3.63 mm). For experiments that required in vivo photostimulation, optical fibers were implanted bilaterally at −10° approximately 200 μm over the dorsal BNST (relative to bregma: ML ±0.90 mm, AP 0.23 mm, DV −4.15 mm) and secured with a dental cement headcap. After surgery, mice were given Tylenol water and allowed to recover for 3 weeks or longer before starting experiments.

Behavioral Testing

Chemogenetics

Behavioral testing was initiated 3 weeks after viral injection of Gq-coupled designer receptors exclusively activated by designer drugs (Gq DREADD: a modified human M3 muscarinic receptor [hM3Dq]) or the control fluorophore. 3 mg/kg CNO or vehicle was administered with an intraperitoneal (i.p.) injection 30 minutes prior to testing to ensure activation of Gq signaling in vlPAG/DR^DA+ neurons during the behavioral assay. CNO and vehicle administration were counterbalanced by day to account for the order of drug and test exposure, with at least one day between each test session.

Optogenetics

Behavioral testing was initiated 6–8 weeks after viral injection of ChR2 or the control fluorophore. Subjects were acclimated to fiber-optic patch cord (Doric Lenses Inc, Quebec, Canada) tethering three days prior to testing. During behavioral testing, mice with optical fibers implanted over the BNST received 473 nm photostimulation with 5-ms pulses at 20 Hz and 10–15 mW power for varying lengths depending on the assay.

Laser Protocol by Assay:

-Hargreaves: Every two trials alternate by Laser OFF and ON sessions (i.e. OFF: T1-T2,T5-T6, ON: T3-T4,T7-T8). For each “ON” trial, the laser is turned on for one minute prior to heat exposure and remains on until the hind paw withdraws from the heat source. Due to sustained anti-nociceptive effects following the initial Laser ON sessions (T3-T4), analysis was restricted T1-T4.

-Von Frey: The first day serves as a Laser OFF session, where subjects are tethered to the patch cable without any input from the laser. The second day is a Laser ON session, where the laser is on for the duration of the test.

-Acetic Acid + Nesting: The laser is turned on for two 15-minute epochs of a 60-minute test. The first and fourth epochs are Laser OFF sessions, while the second and third epochs are Laser ON sessions.

-Real-Time Place Preference/Aversion: The laser switches on whenever the subject is on the designated stimulation side. Conditional activation of the laser based on subject location is valid for the duration of the test.

-Tail Immersion: A baseline measure is taken in the Laser OFF session. This is then followed by the Laser ON session an hour later, where the laser is on for the duration of the test.

-Sociability Test: Laser OFF and ON sessions are separated between two test days, where the assignment of laser status is counterbalanced for each day. The laser is on for the duration of the test during Laser ON sessions.

CRISPR/Cas9

Behavioral testing was initiated 3 weeks after viral injection of a 1:3 mix of Cre-inducible hM3Dq and TH CRISPR or CTRL. Viruses associated with TH deletion were provided by Dr. Larry Zweifel, with previous validation experiments detailed in Hunker et al. (2020). The same drug administration and counterbalancing parameters were maintained throughout all chemogenetic experiments (see Chemogenetics section). Assessment of systemic morphine administration consisted of a 3 mg/kg subcutaneous [s.c.] injection 30 minutes prior to testing.

Inflammatory Pain Model

To model inflammatory pain, subjects were given a 50 μl intraplantar injection of Complete Freund’s Adjuvant (CFA; Sigma, St. Louis, MO), an antigen consisting of heat-killed Mycobacterium tuberculosum, in the plantar surface of a single hind paw. Behavioral testing and slice physiology experiments were initiated three days after paw injections, around the time that CFA exhibits maximum inflammatory hyperalgesia.

Hargreaves

The Hargreaves test was used to measure thermal nociceptive sensitivity (Hargreaves et al., 1988). Subjects were placed in Plexiglas boxes on an elevated glass surface and habituated to the behavioral apparatus for a minimum of 30 minutes. The mid-plantar surface of each hind paw was then exposed to a series of heat trials with 10-minute inter-trial intervals. Trials were conducted with radiant heat exposures sequentially alternating between left and right paws. Six trials were conducted for all experiments except those using optogenetics, which required eight trials to balance light exposure sessions. Beam intensity was set to 25 on the IITC Plantar Analgesia Meter (IITC Life Science, Woodland Hills, CA), producing basal paw withdrawal latencies of approximately 4–6 seconds. A cutoff time of 20 seconds was set to prevent excessive tissue damage.

Von Frey

The Von Frey test was used to measure mechanical nociceptive sensitivity. Subjects were confined to Plexiglas boxes on a custom-made elevated metal wire surface (90 × 20 × 30 cm) and habituated to the behavioral apparatus for a minimum of 30 minutes. Nylon monofilaments of forces ranging from 0.008 to 2 grams (g) were applied to the hind paw using the simplified up-down method (SUDO) described in Bonin et al. (2014). Starting with a mid-range force (0.16 g), the filament was applied to the mid plantar surface of the hind paw for ten trials, then repeated with ascending or descending forces depending on the number of paw withdrawals. Withdrawal thresholds were defined as the minimum force (g) filament that elicits a withdrawal reflex for ≥50% of the trials.

Tail Immersion

The tail immersion test was used to measure the tail flick response following exposure to a water-based thermal nociceptive stimulus. Subjects were restrained in Wypall fold wipers (Kimberly-Clark, Irving, TX) and tails were exposed to 50° C water in the test apparatus (Isotemp 110 Water Bath; Fisher Scientific, Hampton, NH). The tail flick latency was measured in two consecutive trials, where readings were taken 1 cm apart on the tail and averaged together. A cutoff time of 10 seconds was set to minimize tissue damage.

Open Field

Subjects were placed into a white Plexiglas open field (50 × 50 × 25 cm) and allowed to freely explore the arena for 10 minutes. The center of the open field was defined as the central 25% of the arena, where light levels were approximately at 30 lux. Tracking of subject location and activity was achieved with EthoVision (Noldus Information Technologies, Wageningen, Netherlands).

Acetic Acid + Nesting

Nesting procedures were adapted from Negus et al. (2015). The home cages of singly housed mice were transferred from the colony room to a sound attenuated testing room, where subjects habituated for 30 minutes. To model pain-related functional impairments, 10 mg/ml injections (i.p) of ddH2O (Day #1) or 1% acetic acid (AA; Day #2) were administered prior to testing. Nestlets were then divided into six equally sized pieces and distributed into individual zones within the home cage. After 60 minutes, nesting activity was measured by the number of zones cleared. Following treatment with AA, visceral nociceptive sensitivity was determined by handscoring writhing behavior (i.e. abdomen contraction with hind limb extension [Bagdas et al. 2016]). Locomotor activity was tracked during ddH2O and AA sessions using EthoVision.

Real-Time Place Preference/Aversion

Subjects were placed in a custom-made black Plexiglas apparatus (52 × 26 × 26 cm) where counterbalanced sides of the arena were assigned for laser stimulation and non-stimulation. Exploration of the stimulation side resulted in 20-Hz laser stimulation, indicative of the rewarding or aversive properties driven by light-sensitive neuronal populations. Mice were able to freely explore the apparatus for 20 minutes, with activity tracking and laser status conducted through EthoVision.

Sociability Test

The three chamber sociability test was used to compare preferences for interaction with a social target versus an object (Moy et al., 2004). The apparatus consisted of a rectangular Plexiglas box (62 × 37 × 22 cm) containing a central compartment and two end compartments. A 10” diameter wire cup was positioned in the top corner of each end compartment. Subjects were placed in the central compartment and allowed to habituate to the apparatus for 10 minutes. In a subsequent session, a conspecific adult mouse (~6–8 weeks of age) and object (colored labelling tape [Fisher Scientific, Hampton, NH]) were placed in the wire cups to form a social chamber on one end, and a non-social chamber on the other. Subjects were then left to explore the apparatus for 10 minutes. Chamber designations were randomly assigned and counterbalanced. All trials were recorded and tracked with EthoVision to determine the time spent in the social and non-social chambers, as well as the time spent exploring the conspecific mouse versus the object. The ratio of interaction time with each stimulus was then used as a measure of sociability. Subjects were first tested with male conspecific mice, then female conspecific mice, allowing for sociability and locomotion to be measured and categorized for same and opposite sex conspecific mice. After 24 hours of single housing (i.e. social isolation [Matthews et al., 2016]), subjects were tested again with male conspecific mice.

Locomotion

Subjects were tested in a locomotor activity chamber (Accuscan Instruments, Columbus, OH). Following a 30-minute habituation, mice were treated with an i.p. injection of saline or CNO / s.c. injection of saline or morphine, then returned to the apparatus, where they freely explored the chamber for an additional 90 minutes. Locomotor measures including distance traveled, velocity, and % mobility were tracked with EthoVision.

Physiology

Brain Slice Preparation

Brains were collected 6–12 weeks after infusion of AAV5-EF1α-DIO-ChR2-eYFP in the vlPAG/DR of male and female TH-Cre mice. Subjects were deeply anesthetized with isoflurane for rapid decapitation and tissue retrieval. Brains were immediately sectioned on a Leica 1200S vibratome (Leica Microsystems, Wetzlar, Germany), where coronal slices of the BNST (300 μm) were collected and transferred to a 30 ± 1°C Isotemp 110 Water Bath (Fisher Scientific, Hampton, NH) containing chambers submerged in oxygenated artificial cerebral spinal fluid (ACSF; in mM:124 NaCl, 4.4 KCl, 2 CaCl₂, 1.2 MgSO₄, 1 NaH₂PO₄, 10.0 glucose, and 26.0 NaHCO₃). Following an hour of incubation, slices were placed in a recording chamber (Warner Instruments) containing oxygenated ACSF (30°C) that flowed through at rate of 2 ml/min. vlPAG/DR^DA+ terminals and BNST neurons were visualized using infrared differential interference contrast (DIC) video-enhanced microscopy (Olympus) and a 470 nm fluorescent LED illumination system (CoolLED, Andover, NH).

Slice Whole-Cell Electrophysiology

In TH-Cre mice, fluorescently labeled terminals expressing ChR2 in the BNST were visualized and stimulated with a 470 nm LED. The following stimulation protocols were used to assess optically-evoked excitatory and inhibitory transmission between vlPAG/DR and BNST neurons:

Voltage clamp:

(i) Single 1-ms Pulse at −55 mV, (ii) Single 1-ms Pulse at +10 mV, (iii) Paired-Pulse Ratio at −55 mV with inter-sweep intervals (ISI) of 50, 100, 150, 200, and 250 ms, (iv) Paired-Pulse Ratio at +10 mV with ISI of 50, 100, 150, 200, and 250 ms. All experiments were conducted with slices in ACSF. A cesium-methanesulfonate-based internal solution (in mM: 135 cesium methanesulfonate, 10 KCl, 10 HEPES, 1 MgCl₂, 0.2 EGTA, 2 QX-314, 4 MgATP, 0.3 GTP, 20 phosphocreatine, pH 7.3, 285–290 mOsmol) was used for all voltage clamp experiments. In a subset of these experiments, tetrodotoxin (TTX, 10 μM) and 4-aminopyridine (4-AP) was bath applied for 10 minutes to isolate monosynaptic currents.

Current clamp:

20 Hz (5-ms pulse width, 20 pulses). All experiments were conducted with slices in ACSF, then ACSF + 3 mM Kynurenic Acid + 25 μM Picrotoxin, then ACSF + 3 mM Kynurenic Acid + 25 μM Picrotoxin + 2 μM SCH-23390 and/or 2 μM Sulpiride, with a ≥10-minute wash-on period allotted prior to repeating the protocol in a new external solution. A potassium gluconate-based internal solution (in mM: 135 K+ gluconate, 5 NaCl, 2 MgCl₂, 10 HEPES, 0.6 EGTA, 4 Na₂ATP, 0.4 Na₂GTP, pH 7.3, 285–290 mOsmol) was used for all current clamp experiments. Recordings acquired with a Multiclamp 700B amplifier were digitized at 10 kHz, filtered at 3 kHz, and analyzed with Clampfit 10.7 (Molecular Devices, Sunnyvale, CA, USA). Borosilicate glass capillaries were pulled using a Flaming-Brown micropipette puller (Sutter Instruments, Novato, CA) to obtain pipette resistances ranging from 2 to 5 MΩ. Input and access resistance were monitored throughout experiments, with ≥20% changes in access resistance excluded for data analysis.

Fast-Scan Cyclic Voltammetry

DA detection was achieved by positioning a carbon fiber microelectrodes (CFME; made inhouse) in the dorsal BNST, applying a potential of −0.4 V (vs Ag/AgCl), then rapidly ramping up to 1.3 V (at 400 V/sec) at a rate of 10 Hz (Tarheel CV, Labview; National Instruments, Austin, TX). Fluorescently labeled terminals expressing ChR2 in the BNST were then stimulated with light pulses from a 437 nm LED. Stimulation protocols varied by pulse (1 pulse, 20 Hz with 5, 10, 20, and 40 pulses [5-ms pulse width, 1 mW]) and frequency (2, 5, 10, 20, and 30 Hz with 20 pulses [5-ms pulse width, 1 mW]). Background subtracted cyclic voltammograms (CVs) were then analyzed using HDCV (UNC Chapel Hill). In a subset of these experiments, 2 μM sulpiride was bath applied for 10 minutes to assess the effects of D2R antagonism on DA release.

Histology

Immunohistochemistry

Mice were anesthetized with an injection of Avertin (1 ml, i.p.) and transcardially perfused with chilled 0.01 M phosphate-buffered saline (PBS) and 4% paraformaldehyde (PFA) in PBS. Brains were extracted and post-fixed in 4% PFA for 24 hours, then placed in PBS for long-term storage at 4° C. Coronal sections (45 μm thick) of vlPAG/DR, BNST, and other structures of interest were collected using a Leica VT1000S vibratome (Leica Microsystems, Nussloch, Germany) and stored in a 50% glycerol, 50% PBS solution at −4° C until immunohistochemistry was performed. Slices were repeatedly washed for 5-minute cycles in PBS, then permeabilized in a 0.5% Triton X-100/PBS solution for 30 minutes. After a 10-minute PBS wash and 1-hour immersion in blocking solution (0.1% Triton X-100/10% Normal Donkey Serum in PBS), the tissue was incubated overnight in primary antibody diluted in the blocking solution (anti-TH [1:1000]; anti-RFP [1:500]) at 4° C. On the following day, slices went through three 10-minute washes in PBS before being incubated for 2 hours in secondary antibodies diluted in PBS (Alexa Fluor 488 Donkey anti-Mouse [1:200]; Alexa Fluor Cy3 Donkey anti-Rabbit [1:200]). The tissue was then washed in PBS for four 10-minute cycles. Processed slices were mounted on slides and coverslipped using Vecta-Shield Mounting Medium with DAPI (Vector Laboratories, Burlingame, CA) in preparation for imaging.

In Situ Hybridization

Following isoflurane anesthetization and rapid decapitation, the brains of C57BL/6J mice (6–8 weeks; Jackson Laboratory, Bar Harbor, ME) were collected and placed on aluminum foil, where they were immediately frozen on dry ice and stored in a −80° C freezer. Using a Leica CM3050 S cryostat (Leica Microsystems, Wetzlar, Germany), coronal sections of BNST (12 μM thickness) were obtained and directly mounted onto Superfrost Plus slides (Fisher Scientific, Hampton, NH), then kept at −80° C. In order to fluorescently label Drd1a and Drd2 mRNA in the BNST, slices were preprocessed with 4% PFA and protease reagent, incubated with target probes for mouse Drd1a and Drd2, then fluorescently labeled with probes targeting the corresponding channels of each receptor (Drd1a in 550, Drd2 in 647; Advanced Cell Diagnostics, Newark, CA). The processed slides were then covered using Vecta-Shield Mounting Medium with DAPI in preparation for imaging.

Confocal Microscopy

All fluorescent images were acquired with the Zeiss 800 Upright confocal microscope and ZenBlue software (Carl Zeiss AG, Oberkochen, Germany), with equipment access granted through the Hooker Imaging Core at UNC Chapel Hill. Validation of virus expression/injection site, optical fiber placement, and immunoreactivity were accomplished with tiled and serial z-stack images obtained through a 20x objective (2 μm optical slice thickness). Images were processed in FIJI (Schindelin et al., 2012) for manual counting and ZenBlue (Carl Zeiss AG, Oberkochen, Germany) for automated counting. For all histology experiments, the number of cells containing fluorescent puncta was calculated per mm² as a measure of protein or gene expression.

QUANTIFICATION AND STATISTICAL ANALYSIS

Researchers were blinded to genotype/virus treatment for all experiments. Single-variable comparisons were made using paired and unpaired t-tests. Group comparisons were made using one-way ANOVA, two-way ANOVA, or two-way mixed-model ANOVA depending on the number of independent and within-subjects variables in a data set. Following significant interactions or main effects, post-hoc pairwise t-tests were performed and corrected using Sidak’s or Tukey’s post-hoc tests to control for multiple comparisons. Results of statistical testing are reported in the Supplemental Information with significance indicated through markers on figures. Data are expressed as mean ± standard error of the mean (SEM), with significance for p values below 0.05 (* p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001). All data were analyzed and visualized with standard statistical software packages from GraphPad Prism 8 (GraphPad Software, San Diego, CA).

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Anti-TH Mouse	Immunostar	Cat #. 22941
488 Donkey Anti-Mouse	Jackson Immunoresearch	Cat #. 715–545-150; RRID: AB 2340846
Anti-RFP Rabbit	Abcam	Cat #. ab62341; RRID: AB 945213
Cy3 Donkey Anti-Rabbit	Jackson Immunoresearch	Cat #. 711–165-152; RRID: AB 2307443
Drd1a	RNAscope/ACD Bio	Cat #. 406491
Drd2	RNAscope/ACD Bio	Cat #. 406501-C3
Bacterial and Virus Strains
AAV8-hSyn-DIO-hM3Dq-mCherry	Krashes et al. (2011)	Cat # 44361; RRID: Addgene 44361
AAV8-hSyn-DIO-mCherry	Bryan L. Roth	Cat # 50459; RRID: Addgene 50459
AAV5-EF1a-DIO-hChR2(H134R)-eYFP	UNC Vector Core	Cat #. AV4313
AAV5-EF1a-DIO-eYFP	UNC Vector Core	Cat #. AV4310
AAV1-CMV-DIO-saCas9-U6-sgTH	Zweifel Lab	Custom Prep
AAV1-CMV-DIO-saCas9-U6-sghTH	Zweifel Lab	Custom Prep
Chemicals, Peptides, and Recombinant Proteins
CNO (Clozapine-N-oxide) dihydrochloride	HelloBio	Cat #. HB6149
CFA (Complete Freund's Adjuvant)	Sigma Aldrich	Cat #. F5881
Acetic Acid	Sigma Aldrich	Cat #. A6283
Kynurenic Acid	Abcam	Cat #. ab120256
Picrotoxin	Abcam	Cat #. ab120315
SCH-23390 hydrochloride	Tocris	Cat #. 0925
(S)-(-)-Sulpiride	Tocris	Cat #. 0895
TTX (Tetrodotoxin citrate)	Abcam	Cat #. ab120055
4-AP (4-aminopyridine)	Abcam	Cat #. ab120122
Experimental Models: Organisms/Strains
TH- Cre mice	Mutant Mouse Regional Resource Centers	Cat #. 029177-UCD; RRID: MMRRC 029177-UCD
C57BL/6J mice	Jackson Laboratory	000664
Software and Algorithms
FIJI	Schindelin et al. (2012)	http://fiji.sc
Prism 8	GraphPad	http://graphpad.com
EthoVision XT 13.0	Noldus Information Technologies	http://noldus.com
Clampfit 10.7	Molecular Devices	http://moleculardevices.com

Highlights.

• vlPAG/DR^DA to BNST activation promotes anti-nociception in male mice

• vlPAG/DR^DA to BNST activation promotes locomotion in female mice

• -Membrane-gated currents driven by optically-evoked DA release differs by sex

• -Dopamine is required for vlPAG/DR driven anti-nociception