Purinergic signaling is required for CO₂/H⁺-induced constriction of RTN arterioles (Hawkins et al., 2017). Therefore, as a first step toward identifying the cellular and molecular basis of this response, we determined the repertoire of P2 receptors expressed by endothelial and smooth muscle cells associated with arterioles in the RTN. We also characterize the P2 receptor expression profile of vascular cells in the caudal aspect of the nucleus of the solitary tract (cNTS) and raphe obscurus (ROb), two putative chemoreceptor regions where P2 receptor signaling does not contribute to respiratory output (Sobrinho et al., 2014). For this experiment, we used Cre-dependent smooth muscle (Myh11^Cre/eGFP) and endothelial cell (Tek^Cre::TdTomato) reporter lines and performed fluorescence activated cell sorting to obtain enriched populations of endothelial or smooth muscle cells (Figure 1—figure supplement 1). Control tissue was obtained at the same level of brainstem but outside the region of interest. Each population of cells was pooled from three to eight adult mice per region for subsequent quantitative RT-PCR for each of the fourteen murine P2 receptor subtypes. The type and relative abundance of P2 receptor transcript expressed by each cell type was similar to the cortex (Vanlandewijck et al., 2018) and relatively consistent between brainstem regions. For example, endothelial cells from each region express ionotropic P2rx4 and P2rx7 and metabotropic P2ry1,2,6,12-14 receptor transcript, albeit at varying levels relative to control (Figure 1A). Also, P2rx1,4 and P2ry2,6,14 receptor transcript were detected in smooth muscle cells from each region and most were under-expressed relative to control (Figure 1B). The P2 transcript profile of vascular cells in the RTN was fairly similar to the ROb (Figure 1A–B) but differed from the cNTS in several ways including lower expression of P2ry6 in endothelial cells (Figure 1A) and higher expression of P2rx1, P2rx4, and P2ry14 in smooth muscle cells (Figure 1B). However, only P2ry2 in the RTN showed a distinct expression profile that was consistent across cell types with a role in vasoconstriction; P2ry2 transcript was detected at lower (−9.2 ± 0.3 log₂ fold change, p=0.0013) and higher (6.8 ± 1.5 log₂ fold change, p=0.0462) levels compared to control in RTN endothelial and vascular smooth muscle cells, respectively (Figure 1A–B). This is interesting because in other brain regions activation of endothelial P2Y₂ receptors favors nitric-oxide-induced vasodilation (Marrelli, 2001), whereas activation of this same receptor in vascular smooth muscle cells mediates vasoconstriction (Brayden et al., 2013; Lewis et al., 2000). These results identify P2Y₂ receptors in vascular smooth muscle cells as potential substrates for CO₂/H⁺-induced ATP-dependent vasoconstriction in the RTN.

To further test this possibility, we used the brain slice preparation, optimized for detecting increases or decreases in vascular tone, to characterize CO₂/H⁺ vascular responses in each chemoreceptor region of interest as well in a non-respiratory region of the somatosensory cortex. For these experiments, we target arterioles based on previously described criteria (Filosa et al., 2004; Mishra et al., 2014). We verified that exposure to CO₂/H⁺ (15% CO₂; pH 6.9) under control conditions decreases the diameter of RTN arterioles by −11.2 ± 4.5% (N = 9 vessels), whereas this same stimulus consistently dilates arterioles in the cNTS (Δ +6.0 ± 2.2%; N = 9 vessels), ROb (Δ +6.5 ± 2.6%; N = 11 vessels) and somatosensory cortex (Δ +3.7 ± 0.9%; N = 9 vessels) (F_3,34=13.49; p<0.0001) (Figure 1Ci–Cii). Consistent with a requisite role of P2Y₂ receptors in this response, we found that CO₂/H⁺-induced constriction of RTN arterioles was eliminated by blockade of P2Y₂ receptors with AR-C118925 (10 μM) (Δ +0.2 ± 0.3%; N = 6 vessels) (T₅ = 1.223; p>0.05) (Figure 1Di–Dii) and mimicked by application of a selective P2Y₂ receptor agonist (PSB 1114; 200 nM) (Δ −2.3 ± 0.7%; N = 11 vessels) (T₁₀ = 3.460; p<0.010) (Figure 1Ei–Eii). As a negative control, we confirmed that PSB 1114 (200 nM) minimally affects vascular tone in the cNTS (Δ −0.05 ± 0.09%; N = 6 vessels; T₅ = 0.3898; p>0.05) and ROb (Δ −0.07 ± 0.19%; N = 7 vessels; T₆ = 0.6076; p>0.05) (Figure 1Ei–Eii), where P2Y₂ receptors were detected at low levels in both endothelial and smooth muscle cells (Figure 1A–B). We also confirmed that the response of RTN arterioles to CO₂/H⁺ was retained when neural action potentials were blocked by bath application of TTX (0.5 μM) (Figure 1—figure supplement 2), indicating that CO₂/H⁺ constriction of RTN arterioles differs from canonical neurovascular coupling both in terms of polarity and dependence (or lack thereof) on neuronal activity (Ladecola, 2017). Also, despite evidence that extracellular ATP can be rapidly metabolized to adenosine (a potent vasodilator including in the RTN [Hawkins et al., 2017]) the response of RTN arterioles to CO₂/H⁺ was unaffected by 8-phenyltheophylline (8 PT; 10 μM) to block adenosine A1 receptors or sodium metatungstate (POM 1; 100 µM) to inhibit ectonucleotidase activity (Figure 1—figure supplement 2). Furthermore, exposure to CO₂/H⁺ has been shown to trigger prostaglandin E₂ (PGE₂) release that may contribute to RTN chemoreception by a mechanism involving EP₃ receptors (Forsberg et al., 2016); therefore, we also tested for a role of PGE₂/EP₃ signaling in CO₂/H⁺-dependent regulation of RTN arteriole tone. We found bath application of PGE₂ (1 µM) decreased RTN arteriole tone by −2.7 ± 0.4% (T₉ = 2.787, p=0.0212, data not shown). However, CO₂/H⁺ constriction of RTN vessels was retained in the presence of L-798,106 (0.5 µM) to block EP₃ receptors (Δ −9.6 ± 4.0%, T₇ = 2.675, p=0.0318) (Figure 1—figure supplement 2), suggesting PGE₂/EP₃ signaling is dispensable for CO₂/H⁺ regulation of arteriole tone in the RTN.

Astrocytes are thought to contribute to CO₂/H⁺induced vasodilation in the cortex (Howarth et al., 2017) and vasoconstriction in the RTN (Hawkins et al., 2017). Consistent with this, we found that exposure to t-ACPD (50 µM), an mGluR agonist widely used to elicit Ca²⁺ elevations in astrocytes (Howarth et al., 2017), caused constriction of RTN arterioles (Δ −5.9 ± 1.7%; N = 8 vessels) and dilation of arterioles in the cNTS (Δ +7.6 ± 2.5%; N = 8 vessels), ROb (Δ +6.8 ± 1.9%; N = 8 vessels) and somatosensory cortex (Δ +2.9 ± 0.8%; N = 5 vessels) (F_3,25=10.18; p<0.0001) (Figure 1Fi–Fii). These results show that CO₂/H⁺-dependent regulation of vascular tone and roles of astrocytes in this process are fundamentally different in the RTN compared to other brain regions, and point to differential purinergic signaling mechanisms downstream of astrocyte activation.

To determine whether P2Y₂ receptors in the RTN regulate CO₂/H⁺ vascular reactivity in vivo, we measured the diameter of pial vessels on the ventral medullary surface (VMS) in the region of the RTN during exposure to high CO₂ under control conditions (saline) and when P2Y₂ receptor are blocked with AR-C118925. Consistent with our slice data, we found under saline control conditions that increasing end-expiratory CO₂ to 9–10%, which corresponds with an arterial pH of 7.2 pH units (Guyenet et al., 2005), constricted VMS vessels by −7.8 ± 0.7% (p=0.01, N = 7 animals) (Figure 2A). However, when P2Y₂-receptors are blocked by application of AR-C118925 (10 µM) to the VMS, increasing inspired CO₂ resulted in a vasodilation of 5.2 ± 0.9% (Figure 2A) (p=0.05; N = 7 animals). Thus, in the absence of functional P2Y₂ receptors, RTN vessels respond to CO₂/H⁺ in a manner similar to other brain regions. Also consistent with our slice data, we found that activation of P2Y₂ receptors by application of PSB1114 (100 µM) to the VMS constricted vessels in the region of the RTN by −6.3 ± 0.9% (p=0.05, N = 7 animals) (Figure 2B).

To correlate P2Y₂-dependent vasoconstriction in the RTN region with respiratory behavior, we simultaneously measured systemic blood pressure and external intercostal electromyogram (Int_EMG) activity (as a measure of respiratory activity) in urethane-anesthetized mice during exposure to low and high CO₂ following VMS application of saline or AR-C118925. We found that VMS application of AR-C118925 (1 mM) minimally affected the CO₂/H⁺ apneic threshold (3.3 ± 0.4% vs. saline: 3.1 ± 0.5%) (p>0.05; two-way RM ANOVA; N = 7). However, at high levels of CO₂ (9–10% etCO₂) AR-C118925 decreased intercostal EMG amplitude by 23 ± 5% (F_2,74 = 69.83; p=0.021; N = 7 animals) (Figure 2C,E), but with no change in frequency (F_2,74 = 1.09; p>0.05; N = 7 animals) (Figure 2C,F). Conversely, VMS application of PSB1114 (100 µM) while holding etCO₂ constant at 5% increased intercostal EMG amplitude by 20 ± 3% (F_2,74 = 96.14; p=0.02; N = 7 animals) (Figure 2D,H), again with no change in frequency (F_2,74 = 0.79; p>0.05; N = 7 animals) (Figure 2D,I). These treatments had negligible effects on systemic mean arterial pressure (MAP) (AR-C118925: 113 ± 5; PSB1114: 112 ± 5; saline: 114 ± 4 mmHg; F_2,74 = 1.29; p>0.05; N = 7 animals) (Figure 2G,J). These results indicate that P2Y₂ dependent vasoconstriction in the RTN contributes to the drive to breathe.

If the mechanism underlying vascular control of RTN chemoreception involves maintenance of tissue CO₂/H⁺ levels, then by this same logic, vasodilation in the cNTS and ROb should buffer tissue CO₂/H⁺ and limit contributions of these regions to respiratory output. Such a braking mechanism may be important for stabilizing breathing since over-activation of the CO₂/H⁺ chemoreflex is thought to cause unstable periodic breathing (Cherniack and Longobardo, 2006). To test this, we disrupted CO₂/H⁺ dilation by injecting the vasoconstrictor U46619 (a thromboxane A2 receptor agonist) into the cNTS and ROb while measuring cardiorespiratory activity in anesthetized mice breathing a level of CO₂ required to maintain respiratory activity (2–3% CO₂). We found that injections of U46619 (1 μM) into the cNTS alone caused an increase in breathing frequency but together with the ROb resulted in unstable breathing as evidenced by frequent bouts of hyperventilation followed by apnea (Figure 3A) (1.2 ± 0.6, vs. saline: 0.4 ± 0.2 apneas/min; p<0.001) and increased breath to breath variability (Figure 3C–E). Consistent with increased chemoreceptor drive, we found that application of U46619 into these regions lowered the CO₂ apneic threshold from 3.2 ± 0.3% to 2.1 ± 0.1% (p<0.05; two-way RM ANOVA; N = 7) (Figure 3F). Also, consistent with previous work (Yalcin and Savci, 2004), we found that injection of U46619 (1 μM) into the cNTS also increased systemic blood pressure (127 ± 11, vs. saline: 98 ± 4 mmHg), (F_3,65 = 77.62; p<0.01, data not shown), whereas, application of this drug into the ROb alone had negligible effects on cardiorespiratory output. Together, these results show that CO₂/H⁺ induced vasoconstriction in the RTN serves to enhance chemoreceptor drive, while simultaneous CO₂/H⁺-dependent dilation in the cNTS and possibly the ROb limits chemoreceptor activity to stabilize CO₂/H⁺ stimulated respiratory drive.

To definitively test contributions of P2Y₂ receptors in vascular smooth muscle cells to respiratory activity, we created a smooth muscle cell-specific P2Y₂ knockout mouse (P2Y₂ cKO maintained on a C57BL6/J background) by crossing Tagln^Cre (JAX #: 017491) with P2ry2^f/f mice provided by Dr. Cheike Seye (Indiana Univ.). We confirmed that Cre recombinase expression was restricted to vascular smooth muscle cells (Figure 4—source data 1) and that P2ry2 transcript was not detectable in RTN smooth muscle cells from P2Y₂ cKO mice (Figure 4A). Each genotype (Tagln^Cre only, P2ry2^f/f only and P2Y₂ cKO mice) was obtained at the expected ratios and gross motor activity, metabolic activity, blood gases, heart rate and blood pressure were all similar between genotypes and sexes (Table 1, Figure 4—figure supplement 1). Therefore, results from male and female mice of each genotype were pooled. To determine whether P2Y₂ cKO mice exhibit respiratory problems, we used whole-body plethysmography to measure baseline breathing and the ventilatory response to CO₂ in conscious unrestrained adult mice (mixed sex). Consistent with the possibly that smooth muscle P2Y₂ receptor-mediated vasoconstriction in the RTN augments the ventilatory response to CO₂, we found that P2Y₂ cKO mice exhibit normal respiratory activity under room air conditions (Table 1) but pronounced hypoventilation during exposure to graded increases in CO₂ (balance O₂ to minimize input from peripheral chemoreceptors). For example, minute ventilation – the product of frequency and tidal volume – of P2Y₂ cKO mice at 5% and 7% CO₂ was 28% and 35% smaller than control counterparts (Figure 4C–F, Table 2). This chemoreceptor deficit involves a diminished capacity to increase both respiratory frequency (Figure 4D) and tidal volume (Figure 4E) during exposure to CO₂ (Table 2). Similar results were obtained in P2Y₂ cKO mice generated using a different smooth muscle cre line (Myh11^Cre/eGFP) (Figure 4—figure supplement 2). Also, smooth muscle P2Y₂ cKO mice showed a normal ventilatory response to hypoxia (10% O₂; balance N₂, T₇ = 0.1089, p>0.05) (Figure 4—figure supplement 3). We confirmed in vitro that RTN arterioles in slices from P2Y₂ cKO do not respond to PBS 1114 (0.4 ± 0.1%, T₇ = 1661, p>0.05, N = 8 vessels). Interestingly, we also found that RTN arterioles in slices from P2Y₂ cKO mice fail to constrict in response to CO₂/H⁺ or tACPD, but rather respond in a manner similar to arterioles in other brain regions. For example, both CO₂/H⁺ and tACPD dilated RTN (CO₂/H⁺: 4.7 ± 0.2%, T₉ = 3.312, p=0.009, N = 10 vessels; tACPD: 3.2 ± 0.7%, T₇ = 2.864, p=0.035, N = 8 vessels) and cortical (CO₂/H⁺: 7.3 ± 1.5%, T₆ = 4.307, p=0.0051, N = 7 vessels; tACPD: 6.7 ± 0.9%, T₆ = 2.739, p=0.0338, N = 7 vessels) arterioles in slices from P2Y₂ cKO mice (Figure 4—figure supplement 4). This suggests CO₂/H⁺ induced vasodilation is a general phenomenon mediated by undetermined mechanisms, but in the RTN is response is countered by smooth muscle P2Y₂ receptor-mediated constriction.

We next tested whether targeted re-expression of P2Y₂ in RTN smooth muscle cells would rescue the blunted chemoreflex in P2Y₂ cKO mice. An adeno-associated virus (AAV2-Myh11p-eGFP-2A-mP2ry2, Vector Biolabs) was injected bilaterally into the RTN of adult P2Y₂ cKO (Tagln^Cre::P2ry2^f/f) to drive P2Y₂ expression in smooth muscle cells. Two weeks after virus injection into the RTN region, we confirmed that ~80% of smooth muscle α2 actin (Acta2)-immunoreactive cells in the RTN are also GFP⁺ (Figure 4—figure supplement 5). Some background eGFP labeling was observed, but not associated with DAPI labeling (Figure 4—figure supplement 5). We also confirmed RTN smooth muscle cells from P2Y₂ rescue mice show increased P2ry2 transcript levels compared to P2Y₂ cKO (p=0.0073) but not to the same level as cells from control mice (H₃ = 7.2, p=0.0219) (Figure 4A). Importantly, re-expression of P2Y₂ receptors in RTN smooth muscle cells resulted in a full rescue of the minute ventilatory response to CO₂ (Figure 4C–F, Table 2) (F_1,7=17.75, p=0.004, N = 13 animals). Conversely, bilateral RTN injections of control virus (AAV2-Myh11p-eGFP, Vector Biolabs) minimally affected respiratory parameters of interest in P2Y₂ cKO mice (Figure 4—figure supplement 2). Furthermore, re-expression of P2Y₂ receptors in RTN smooth muscle cells also rescues CO₂/H⁺ vascular reactivity. Specifically, we found that, in slices from P2Y₂ cKO mice, RTN arterioles transfected with AAV showed a robust constriction in response to CO₂ (−19.9 ± 6.9%, p=0.0421), tACPD (−6.2 ± 1.9%, p=0.0092) and PSB1114 (−9.3 ± 4.9%, p=0.0154) (Figure 4—figure supplement 4). These results identify the first vascular element of respiratory control by showing that P2Y₂ receptors in RTN vascular smooth muscle cells are required for the normal ventilatory response to CO₂.

There are several limitations of this work that should be recognized. First, our in vitro experiments utilized the slice preparation because it allows for easy visualization of vessels in each region of interest and pharmacological manipulation of candidate mechanisms independent of potential confounding effects of blood pressure on myogenic tone. However, because blood vessels in brain slice have limited myogenic tone these experiments were performed in the presence of U-46619 to pre-constrict vessels ~ 30% (Blanco et al., 2008). We also used a large stimulus (15% CO₂) to characterize CO₂/H⁺ vascular reactivity in vitro. For these reasons, we also confirmed in vivo in the absence of U-46619 that more physiological levels of CO₂ (9–10%) constricted arterioles in the RTN by a P2Y₂-dependent mechanism. Second, we and others (Yalcin and Savci, 2004) showed that administration of U46619 into the NTS increased blood pressure. Although baroreceptor activation has a fairly mild effect on respiratory activity (McMullan et al., 2009), it remains possible that the baroreflex contributed to respiratory instability observed during this experimental manipulation. Furthermore, neurons and astrocytes may express thromboxane A2 receptors (Gao et al., 1997; Nakahata et al., 1992) so it is possible respiratory instability caused by injections of U46619 into the cNTS and ROb may involve direct neural or astrocyte activation. Third, CO₂/H⁺ typically increases minute ventilation by increasing both rate and depth of breathing (Souza et al., 2019; Takakura et al., 2014), yet ventral surface application of drugs to manipulate P2Y₂ receptors only affected tidal volume (Figure 2E–J). Therefore, it is possible other nearby respiratory centers were affected in this experiment. However, this issue is mitigated by more targeted genetic approaches showing that deletion of P2Y₂ from smooth muscle cells blunted the rate and depth of respiratory responses to CO₂, and re-expression of P2Y₂ receptors only in RTN smooth muscle cells fully rescued the CO₂/H⁺ chemoreflex.

Blood flow in the brain is controlled by several mechanisms including neural activity, which leads to vasodilation and increased blood flow by a process termed neurovascular coupling (Ladecola, 2017). Autoregulation is a mechanism that maintains cerebral blood flow in response to changes in perfusion pressure (Paulson et al., 1990). For example, myogenic tone increases with increased perfusion pressure and relaxes with decreased pressure. CO₂/H⁺ also functions as a potent vasodilator in most brain regions (Hoiland et al., 2019) and is perhaps the most potent regulator of cerebrovascular tone since high CO₂/H⁺ can impair both autoregulation (Ogoh and Ainslie, 2009; Perry et al., 2014) and neurovascular coupling (Maggio et al., 2013; Maggio et al., 2014). Since CO₂/H⁺ is a waste product of metabolism, CO₂/H⁺ vascular reactivity provides a means of matching local blood flow with tissue metabolic needs. However, at the level of the RTN this need appears secondary to chemoreceptor regulation of respiratory activity. For example, we (Hawkins et al., 2017) and others (Wenzel et al., 2020) showed that exposure to high CO₂/H⁺ constricted RTN arterioles, and disruption of this response blunted the ventilatory response to CO₂.

The cellular and transmitter basis for CO₂/H⁺-induced constriction of RTN arterioles appears to involve CO₂/H⁺-evoked ATP release from local astrocytes. For example, previous work showed that (i) CO₂/H⁺ caused discrete release of ATP near the RTN region (Gourine et al., 2005); (ii) astrocytes in the RTN show Ca²⁺ responses to mild acidification (Gourine et al., 2010); and (iii) activation of astrocytes with tACPD, an mGluR agonist widely used to elicit Ca²⁺ elevations in astrocytes (Howarth et al., 2017), constricted RTN arterioles by a purinergic-dependent mechanism (Hawkins et al., 2017; Figure 2B), whereas application of an ATP receptor blocker to the ventral surface blunted CO₂/H⁺-induced vasoconstriction and the ventilatory response to CO₂ (18). Here, we extend this work by showing that CO₂/H⁺ vascular responses in the RTN are opposite to other chemoreceptor regions, and we identify P2Y₂ receptors in vascular smooth muscle cells as requisite determinants of RTN CO₂/H⁺ vascular reactivity and chemoreception. For example, P2Y₂ is a metabotropic receptor that couples with Gq, Go and G12 proteins (Erb and Weisman, 2012; activation of this receptor in smooth muscle is associated with vasoconstriction; Brayden et al., 2013; Lewis et al., 2000), whereas its activation in endothelial cells favors nitric oxide-mediated vasodilation (Marrelli, 2001). Consistent with a role in vasoconstriction, we found that P2Y₂ transcript was expressed at higher than control levels in RTN smooth muscle cells and at lower than control levels in RTN endothelial cells (Figure 1A–B). Furthermore, pharmacological experiments show in vitro (Figure 1C–E) and in anesthetized mice (Figure 2A–B) that CO₂/H⁺-induced constriction of RTN arterioles was eliminated by blocking P2Y₂ receptors with AR-C118925 and mimicked by a selective P2Y₂ receptor agonist (PSB 1114). We also show in awake mice that genetic deletion of P2Y₂ from smooth muscle cells blunted the ventilatory response to CO₂, and re-expression of P2Y₂ receptors only in RTN smooth muscle cells fully rescued the CO₂/H⁺ chemoreflex (Figure 4A–F). These results establish P2Y₂ receptors in RTN smooth muscle cells as requisite determinants of respiratory chemoreception.

It should be noted that other cell types and signaling pathways may also contribute to RTN vascular reactivity. For example, endothelial cells including those in the RTN, express H⁺ activated GPR4 receptors that when activated facilitates release of vasoactive effectors including thromboxane to mediate vasoconstriction (Wenzel et al., 2020). Consistent with this, disruption of GPR4 signaling blunted CO₂/H⁺ vascular reactivity in the RTN and the ventilatory response to CO₂ (Wenzel et al., 2020). Although untested, it is also possible RTN endothelial cells release other vasoactive signals including ATP, and so may contribute to CO₂/H⁺-evoked purinergic-dependent constriction of arterioles in this region. Prostaglandin E₂ (PGE₂) is also thought to contribute to CO₂/H⁺ vascular reactivity. For example, in the cortex CO₂/H⁺ facilitates Ca²⁺ oscillations in astrocytes (Gourine et al., 2010; Howarth et al., 2017) leading to enhanced arachidonic acid metabolism and release of PGE₂, which triggered arteriole dilation by activation of prostaglandin (EP₁) receptors (Howarth et al., 2017). In the RTN, CO₂/H⁺ also causes PGE₂ release most likely from astrocytes (Forsberg et al., 2017) and contributes to RTN chemoreception by a mechanism involving EP3 receptors (Forsberg et al., 2016). However, selective blockade of EP₃ receptors minimally affected CO₂/H⁺ constriction of RTN arterioles (Figure 1—figure supplement 2), suggesting PGE₂/EP₃ signaling contributes to RTN chemoreception by directly targeting CO₂/H⁺-sensitive neurons or astrocytes.

In contrast to the RTN, we also found that CO₂/H⁺ (Figure 1C) and astrocyte activation by bath application of t-ACPD (Figure 1F) dilatated arterioles in other chemoreceptor regions including the cNTS and ROb. This suggests that CO₂/H⁺-dependent regulation of vascular tone and roles of astrocytes in this process are fundamentally different in the RTN compared to other chemoreceptor regions. Considering vasoconstriction and dilation are expected to increase and decrease tissue CO₂/H⁺ levels respectively, we propose that CO₂/H⁺ constriction in the RTN supports the drive to breathe by ensuring tissue acidification is maintained to stimulate chemoreceptors in this region, whereas CO₂/H⁺ dilation in the cNTS and ROb provides a means of buffering tissue H⁺ and chemoreceptor activity to maintain respiratory stability. Conversely, if CO₂/H⁺ were to cause vasoconstriction across multiple chemoreceptor regions simultaneously, this may exaggerate the ventilatory response to CO₂ to the point of causing unstable periodic breathing (Cherniack and Longobardo, 2006). Indeed, this is thought to be the cause of unstable breathing in patients with global deficits in CO₂/H⁺ vascular reactivity due to congestive heart failure or cerebrovascular disease (Cherniack et al., 2005; Cherniack and Longobardo, 2006). Consistent with this, we show that disruption of CO₂/H⁺ dilatation in the cNTS and ROb by application of the vasoconstrictor U46619 while leaving RTN CO₂/H⁺ constriction unperturbed, increased chemoreceptor drive as evidenced by a decrease in the CO₂/H⁺ apneic threshold and caused unstable periodic breathing during exposure to higher levels of CO₂/H⁺ (Figure 3). However, further work is needed to understand whether and how differential CO₂/H⁺ vascular reactivity in these chemoreceptor regions contributes to breathing problems in disease states.

All procedures were performed in accordance with National Institutes of Health and University of Connecticut Animal Care and Use Guidelines. All experiments used mixed sex C57BL6/J animals at least 3 weeks of age that housed in a 12:12 light dark cycle with normal chow ad libitum. The following transgenic Cre mouse lines were used to perform single cell isolation, FACS, and pooled qPCR: Myh11^Cre/eGFP (Jax Stock #: 007742) and Tek^Cre::TdTomato (Jax Stock #: 008863, 007909), and Tagln^Cre::TdTomato (Jax Stock #: 017491, 007909). Tagln^Cre and Myh11^Cre/eGFP mice were crossed with P2ry2 floxed mice (gifted by Dr. Cheike Seye at Indiana University) to generate double floxed smooth muscle-specific P2Y₂ cKO mice for in vitro and in vivo experimentation.

At least three and no greater than eight, animals (postnatal days P21-P40) were used of either of the following genotypes: Myh11^Cre/eGFP, Tek^Cre::TdTomato, Tagln^Cre::TdTomato, or Tagln^Cre::TdTomato::P2ry2^fl/fl. Animals were euthanized under isoflurane anesthesia and brainstem slices were prepared using a vibratome in ice cold, high sucrose slicing solution containing (in mM): 87 NaCl, 75 sucrose, 25 glucose, 25 NaHCO₃, 1.25 NaH₂PO₄, 2.5 KCl, 7.5 MgCl₂, 0.5 mM CaCl₂ and 5 L-ascorbic acid. Slicing solution was equilibrated with a 5% CO₂-95% O₂ gas mixture before use. Transverse brainstem slices (400 μm thick) were prepared and then immediately enzymatically treated at 34°C with an initial incubation in a papain (6 mg/mL, Sigma) and dithioerythritol (10 mg/mL, ThermoFisher) mixture for 30 min in sucrose dissociation solution containing (in mM): 185 sucrose, 10 glucose, 30 Na₂SO₄, 2 K₂SO₄, 10 HEPES, 0.5 CaCl₂, 6 MgCl₂, 5 L-ascorbic acid, pH 7.4, 320 mOsm, followed by a collagenase IV (10 mg/mL, ThermoFisher) enzyme treatment for 6 min. After enzyme incubation, slices were washed three times in cold dissociation solution and then transferred to an enzyme inhibitor mix containing trypsin inhibitor (10 mg/mL, Sigma), bovine serum albumin (BSA, 10 mg/mL, Sigma), and sodium nitroprusside (SNP, 2 mg/mL, ThermoFisher) in cold sucrose dissociation solution. Shortly thereafter, slices were transferred to a glass Petri dish on ice. Using a plastic transfer pipette and a 15-blade scalpel, each region of interest was microdissected out of the slices and manually separated into sterile microcentrifuge tubes. The control samples were made up of two slices: one slice with the NTS removed and the other with the ROb and RTN removed. Once microdissection is completed, the tissue chunks were warmed to 34°C for 10 min before trituration. A single-cell suspension was achieved by trituration using a 25G and 30G needle sequentially, attached to a 3 mL syringe. Samples were triturated for an average of 5 min. Immediately after, the samples were placed back on ice and filtered through a 30-µm filter (Miltenyi Biotech) into round bottom polystyrene tube for fluorescence-activated cell sorting (FACS).

Animal use was in accordance with guidelines approved by the University of São Paulo Animal Care and Use Committee. A total of 26 adult male C57BL6 (25–28 g) were used for in vivo experiments. General anesthesia was induced with 5% isoflurane in 100% O2. A tracheostomy was made, and the isoflurane concentration was reduced to 1.4–1.5% until the end of surgery. The carotid artery was cannulated (polyethylene tubing, 0.6 mm o.d., 0.3 mm i.d., Scientific Commodities) for measurement of arterial pressure (AP). The jugular vein was cannulated for administration of fluids and drugs. Rats were placed supine onto a stereotaxic apparatus (Type 1760; Harvard Apparatus) on a heating pad and core body temperature was maintained at a minimum of 36.5°C via a thermocouple. The trachea was cannulated. Respiratory flow was monitored via a flow head connected to a transducer (GM Instruments) and CO₂ via a capnograph (CWE, Inc,) connected to the tracheal tube. Paired EMG wire electrodes (AM-System) were inserted into the external intercostal muscle to record respiratory-related activity. After the anterior neck muscles were removed, a basio-occipital craniotomy exposed the ventral medullary surface, and the dura was resected. After bilateral vagotomy, the exposed tissue around the neck and the mylohyoid muscle was covered with mineral oil to prevent drying. Baseline parameters were allowed to stabilize for 30 min prior to recording.

Adult mice (>20 grams) were anesthetized with 3% isoflurane. The right cheek of the animal was shaved, and an incision was made to expose the right marginal mandibular branch of the facial nerve. The animals were then placed in a stereotaxic frame and a bipolar stimulating electrode was placed directly adjacent to the nerve. Animals were maintained on 1.5% isoflurane for the remainder of the surgery. An incision was made to expose the skull and two 1.5 mm holes were drilled left and right of the posterior fontanelle, caudal of the lambdoidal suture. The facial nerve was stimulated using a bipolar stimulating electrode to evoke antidromic field potentials within the facial motor nucleus. In this way, the facial nucleus on the right side of the animal was mapped in the X, Y, and Z direction using a quartz recording electrode.

The viral vector was loaded into a 1.2 mm internal diameter borosilicate glass pipette on a Nanoject III system (Drummond Scientific). Virus was injected at least −0.02 mm ventral to the Z coordinates of the facial nucleus, to ensure injection into the RTN. These same coordinates were used for the left side of the animal. In all mice, incisions were closed with nylon sutures and surgical cyanoacrylate adhesive. Mice were placed on a heated pad until consciousness was regained. Meloxicam was administered 24 and 48 hr postoperatively.

Adult mice (>30 grams) were anesthetized with an induction dose of 3% isoflurane and placed into a sterile field. HD-X11 transmitters (Data Sciences International; DSI) were placed and fixed intraperitoneally with non-absorbable nylon sutures, followed by ECG lead placement on the left and right pectoralis on top of the rib cage. Finally, a pressure catheter was placed into the left carotid artery and threaded through to the aortic arch for continuous blood pressure measurements. Animals were closed with nylon sutures and placed on a heated pad u conscious. Mice were continuously watched for the next three days for any post-operative pain complications while meloxicam was administered 24 and 48 hr postoperatively. Mice were not used for any experimentation until post-operative day 7.

Respiratory activity was measured using a whole-body plethysmograph system (Data Scientific International; DSI), utilizing animal chamber maintained at room temperature and ventilated with air (1.3 L/min) using a small animal bias flow generator. Mice were individually placed into a chamber and allowed 1 hr to acclimate prior to the start of an experiment. Respiratory activity was recorded using Ponemah 5.32 software (DSI) for a period of 15 min in room air followed by exposure to graded increases in CO₂ from 0% to 7% CO₂ (balance O₂). In separate experiments, we characterized the ventilatory response to 10% O₂ (balance N₂). Parameters of interests include respiratory frequency (F_R, breaths per minute), tidal volume (V_T, measured in mL; normalized to body weight and corrected to account for chamber and animal temperature, humidity, and atmospheric pressure), and minute ventilation (V_E, mL/min/g). A 20 s period of relative quiescence after 5 min of exposure to each condition was selected for analysis. All experiments were performed between 9 a.m. and 6 p.m. to minimize potential circadian effects.

Metabolic monitoring (VCO₂, VO₂) was performed using comprehensive lab animal monitoring systems (CLAMS, Columbus Instruments). In short, mice were individually housed on a 12:12 light dark cycle in plastic cages with a running wheel, regular bedding, and regular chow for 1 week before experimentation. Three days before the metabolic experiment, each animal was placed in the CLAMS housing cage with metered water and waste collection. Mice had 2 days to acclimate to the metabolic chamber; on the third day, all results were recorded for a continuous 24 hr period. After data collection, all raw results were exported and averaged out per hour. Then, light and dark periods were determined and averaged per animal for statistical analysis. Both sexes were equally represented in the data set.

Arterial blood gasses were collected from adult mice 6 weeks of age and older (>30 grams). A RAPIDLab 348 blood gas analyzer (Siemens) was used for all blood gas analysis; all calibrations, QC, and use were performed as indicated by the manufacturer. Animals were anesthetized with an induction dose of 3% isoflurane and then quickly switched to 1% isoflurane for the remainder of arterial blood collection. This level of isoflurane minimally affects blood gases (Constantinides et al., 2011; Loeven et al., 2018). The left carotid artery was exposed and quickly cannulated to allow for arterial blood to be collected and analyzed by the blood gas analyzer; no more than 5 s was spent between blood collection and analysis on the blood gas analyzer.

Adult mice were transcardially perfused with 20 mL of room temperature phosphate buffered saline (PBS, pH 7.4) followed by 20 mL of chilled 4% paraformaldehyde (pH 7.4). The brainstem was then removed from the animal and 150 um slices were made using a Zeiss VT100S vibratome. Slices were then incubated in a 0.5% Triton-X/PBS solution for 45 min to permeabilize the tissue. The slices remained in a 0.1% Triton-X/10% Fetal Bovine Serum (FBS, ThermoFisher)/PBS solution for a 12 hr primary antibody incubation of anti-mouse alpha smooth muscle actin (Sigma). The tissue was then washed three times in 0.1% Triton-X/10% FBS/PBS solution; the secondary antibody was incubated with the tissue after the third wash for 2 hr (donkey anti-mouse Alexa Fluor 647, ThermoFisher). The tissue was then washed three times in PBS before mounting on precleaned cover slides with Prolong Diamond with DAPI (ThermoFisher). Imaging of brain slices was achieved with a Leica SP8 confocal microscope.

Data are reported as mean ± SE. Power analysis was used to determine sample size, all data sets were tested for normality using Shapiro-Wilk test, and comparisons were made using t-test, one-way or two-way ANOVA (parametric or non-parametric) followed by multiple comparison tests as appropriate. The specific test used for each comparison is reported in the figure legend and all relevant values used for statistical analysis are included in the results section.