Introduction

Urination is a basic life-maintaining function involving the coordinated control of two functional units of the lower urinary tract (Andersson and K.-E., 2004; Chang et al., 2018; Fowler et al., 2008), namely the bladder detrusor and external urethral sphincter (EUS). The coordination between the bladder and urethra sphincter for initiating or suspending voiding whenever needed should be executed at 100% reliability in healthy subjects. By contrast, impairment of such coordination (Cho et al., 2015; Taweel et al., 2015; Griffiths et al., 2005), at any rate, leads to various lower urinary tract dysfunctions (Sakakibara and Ryuji, 2015; Drake et al., 2014), significantly degrading the quality of life (Milsom et al., 2014; Aoki et al., 2017). While individual neural control pathways to either the bladder detrusor or the urethral sphincter have been extensively studied (Xiao et al., 2021; Lee et al., 2021), little is known about the neural mechanisms underlying their coordination, which impedes progress in developing target-specific therapies for certain urination disorders, such as detrusor-sphincter dyssynergia (DSD) (Stoffel, 2016; Seseke et al., 2019).

Individual neural pathways are present at different levels of the brain for the control of either the bladder or the EUS (Malykhina, 2017; Jin et al., 2020; Mukhopadhyay and Stowers, 2020). For example, at the cortical level, a subset of motor cortex neurons has been found to drive bladder contraction through the projection to the brainstem (Yao et al., 2018). At the brainstem level, the pontine micturition center (PMC; also referred to as Barrington’s nucleus) has long been considered a command hub region for urination control (Benarroch, 2010; Morrison, 2008). More recent studies have demonstrated that neurons expressing corticotropin-releasing hormone (CRH) in the PMC (PMC^CRH+) control bladder contraction (Hou et al., 2016; Ito et al., 2020), and neurons expressing estrogen receptor 1 (ESR1) in the PMC (PMC^ESR1+) relax the sphincter (Keller et al., 2018), respectively. In addition, other subcortical regions, including the periaqueductal gray (PAG) (Rao et al., 2022), lateral hypothalamic area (LHA) (Verstegen et al., 2019; Hyun et al., 2021), and medial preoptic area (MPOA) (Hou et al., 2016), have been suggested to play a potential role in urination control by sending direct projections to the PMC. However, which brain areas coordinate both the bladder and urethra sphincter remains unclear.

The most likely candidate region for coordinated control is the PMC. Neural tracing experiments demonstrate that the PMC directly sends two bundles of glutamatergic axonal projections, one to the sacral parasympathetic nucleus (SPN), where parasympathetic bladder motoneurons are located, which send axons through the pelvic nerves for bladder control; and the other one to the lumbosacral dorsal gray commissure (DGC) interneurons, which inhibit the sphincter motoneurons in the dorsolateral nucleus (DL) for sphincter control via the pudendal nerves (Kawatani et al., 2021; Jin et al., 2020). Early studies show that microinjection of drugs into the PMC or electrical stimulation in the PMC causes bladder contraction, sphincter relaxation, and urination (Mallory, 1991; Noto et al., 1989; Sugaya and De Groat, 1994). However, the PMC consists of molecularly disparate cell subtypes characterized by the expression of different marker genes, e.g. CRH positive cells (Ito et al., 2020; Van Batavia et al., 2021), ESR1 positive cells (Vanderhorst et al., 2005), vesicular glutamate transporter (Vglut2) and vesicular GABA transporter (VGAT) positive cells (Verstegen et al., 2017; Hou et al., 2016). What remains unknown to date is which exact neuronal subpopulation in the PMC accounts for such coordination of the bladder and urethra sphincter. As previous anatomical results have shown that PMC^CRH+ neurons mainly project to the SPN in the spinal cord, whereas PMC^ESR1+ neurons project to both the SPN and DGC in the spinal cord (Keller et al., 2018; Kawatani et al., 2021), we hypothesize that PMC^ESR1+ cells could be a candidate for the coordination control and began our investigation.

Results

Voiding tightly correlates with PMC^ESR1+ cell activity

We performed fiber photometry monitoring of neuronal population Ca²⁺ activity (Rao et al., 2022) (fluorescence indicator: GCaMP6f) in the PMC of awake, unrestrained ESR1-Cre mice (Figure 1A and B, see Methods for detail). Each detected voiding event tightly correlated with a detected Ca²⁺ transient (Ca²⁺ transient preceding voiding by 0.7\0.3-0.9 s, median\25%-75% percentile, same notation hereinafter; n = 260 voiding events from 9 mice; Figure 1C and D), as verified by both an analysis of temporally shuffled data (peak Δf/f value, data: 31.8%\26.7%-33.2%; shuffled: 5.4%\3.7%-6.0%, n = 260 events of 9 mice, p = 3.9e-3, Wilcoxon signed-rank test; Figure 1E) and a blank control with an expression of inert fluorescence indicator, EYFP (Enhanced Yellow Fluorescent Protein), in PMC^ESR1+ cells (detected voiding-associated fluorescence signal events: 100%\100% – 100%, n = 9 mice in test group; 0%\0% – 0%, n = 9 mice in control group, p = 4.1e-5, Wilcoxon rank-sum test; Figure 1F-figure supplement 1). To further test single-neuron correlates of voiding, we performed an ‘opto-tagging’ experiment (Qin et al., 2022) with a tetrode-fiber bundle implanted in the PMC of ESR1-Cre mice injected with AAV2/8-DIO-ChR2-mCherry (Figure 1G; see Methods for detail). Single PMC^ESR1+ units were sorted and tagged by detecting reliable spikes to brief optical stimulation (e.g., latency: 4.04\3.02-4.6 ms, success rate: 97.5%\91.5%-100%; Figure supplement 2). The baseline-corrected firing rate of PMC^ESR1+ units significantly exceeded that of non-PMC^ESR1+ units in the temporal association window with voiding (PMC^ESR1+ units: ‘before’, 0.04\-0.5-0.5 Hz, ‘voiding’, 1.4\-2.5e-11-3.8 Hz, n = 28 cells, p = 0.004; non-PMC^ESR1+ units: ‘before’, 0.8\0.2-1.5 Hz, ‘voiding’, 1\-0.9-1.9 Hz, n = 51 cells, p = 0.6; Wilcoxon signed-rank test; Figure 1H and I). These results suggest that PMC^ESR1+ cells’ firing activities are tightly associated with voiding.

Figure 1.

Additionally, we performed a set of combined physiological monitoring experiments, integrating fiber photometry, cystometry, and electromyography (EMG) of external urethral sphincter (EUS) simultaneously in urethane-anesthetized mice (Figure supplement 3A, see methods for detail). For quantitative measurement, we applied cystometry by continuously infusing saline into the bladder through a microcatheter (30-50 μl/min) to induce regular reflective voiding. A triple correlation of events was routinely observed, i.e., PMC^ESR1+ neuronal activity, bladder pressure elevation, and the bursting pattern of EUS-EMG driving sphincter relaxation (Figure supplement 3B and C). This result was further validated by shuffled data analysis (Figure supplement 3D and E), demonstrating the robustness and time-locking precision of the triple correlation events. However, it is important to note that the association between PMC^ESR1+ cell activity and voiding in the other way around was not always 100%, i.e., for each detected Ca²⁺ transient, there could be either a voiding contraction (VC) or occasionally a non-voiding contraction (NVC) (Biallosterski et al., 2011) (27.1% of all detected events were in such case; Figure supplement 4A-C). Nevertheless, not only was the peak of bladder pressure lower, but also the amplitude of the photometry Ca²⁺ transient, which is known as a reliable report of collective neuronal population activity level, was significantly lower in NVC events than in VC events (NVCs: 8.8%\6.6%-14.7%, n = 62 events from 3 mice; VC: 15.01%\10.3%-26.4%, n = 167 events from 3 mice, p = 6.8e-9, Wilcoxon rank-sum test; figure supplement 4D and E), suggesting that not only the timing but also the strength of PMC^ESR1+ cell activities were tightly correlated with successful voiding.

PMC^ESR1+ cells bidirectionally operate the bladder and sphincter to initiate or suspend voiding

With the tight correlation established, we moved on to test the causal relation between PMC^ESR1+ cell activity and voiding. We started with a ‘loss-of-function’ test in awake mice, i.e., acute photoinhibition of PMC^ESR1+ cells in a closed-loop to trigger the photoinhibition light as soon as the first patch of urine visualized (Figure 2A-figure supplement 5A; see Methods for detail). This test was also accompanied by a blank control in which all experimental conditions were the same except that the inhibitory opsin GtACR1 was absent. In photoinhibition events but not in the blank control events, the urine spot area was significantly reduced (‘Pre’: 34.1\27.1-40.4 cm²; ‘On’: 9.9\7.9-11.1 cm², n = 12 mice in test group, p = 4.9e-4; ‘Pre’: 35.2\27.9-43.9 cm²; ‘On’: 38.9\29.98-43.6 cm², n = 8 mice in control group, p = 0.3; Wilcoxon signed-rank test; Figure 2B and C), and the urination duration was significantly reduced as well (‘Pre’: 5.6\5.1-6.6 s; ‘On’: 1.38\1.19-1.42 s, p = 4.9e-4, Wilcoxon signed-rank test; Figure 2D and E). Effectively, the ongoing urination was fully suspended after a latency of 0.3 ± 0.1 sec from the onset of photoinhibition (Figure supplement 5B). To understand the physiological process during photoinhibition of PMC^ESR1+ cells, we performed a simultaneous recording by measuring both the bladder pressure and the electromyograph of the external urethral sphincter (EUS-EMG) under urethane anesthesia (Figure 2F; see Methods for detail). This experiment also involved a closed-loop operation to trigger the light when observing the onset of the phasic bursting activity of the EUS-EMG that is known to be directly associated with successful voiding in rodents (Kadekawa et al., 2016a; Langdale and Grill, 2016). The maximum relative change in bladder pressure (ΔPressure) upon voiding event was significantly reduced during photoinhibition compared to pre-inhibition baseline events (‘Pre’: 11.6\10.4-15.4 cm H₂O, ‘On’: 8.1\5.9-9.4 cm H₂O, n = 8 mice, p = 7.8e-3, Wilcoxon signed-rank test), but no such reduction was observed in the blank control group (‘Pre’: 10.7\6.9-12.03 cm H₂O, ‘On’: 10.5\7.2-13.4 cm H₂O, n = 7 mice, p = 0.2, Wilcoxon signed-rank test; Figure 2G and H). In the meanwhile, photoinhibition of the PMC^ESR1+ cells also halted the voiding-associated, phasic bursting activity of EMG, effectively reducing the sphincter bursting duration (‘Pre’: 4.1\3.2-4.9 s, ‘On’: 0.98\0.9-1.2 s, p = 7.8e-3, Wilcoxon signed-rank test; Figure 2G and I) at a latency of 0.1 ± 0.02 sec (Figure supplement 5C and D), but no such effect was observed in the blank control group (‘Pre’: 3.3\2.9-4.8 s, ‘On’: 3.4\2.96-5.4 s, p = 0.8, Wilcoxon signed-rank test). Furthermore, the immediate suspension of an ongoing voiding event by photoinhibition of PMC^ESR1+ cells resulted in an expected side effect that the threshold of bladder pressure to initiate the next voiding event (post-photoinhibition) became higher (figure supplement 5E). A control test for the above set of ‘loss-of-function’ experiments was to test whether shorter durations of photoinhibition also had the same urination suspension effect. To address this, we performed additional sets of control experiments in which a shorter duration of photoinhibition (5 s light-on, instead of 60 s) resulted in the same, 100% suspension effect (Figure supplement 6). These data together reveal that the acute photoinhibition of PMC^ESR1+ cells halted both bladder contraction and sphincter relaxation, thereby leading to a full suspension of the ongoing voiding process.

Figure 2.

Then, we performed a ‘gain-of-function’ test in awake mice, i.e., acute photoactivation of PMC^ESR1+ cells that expressed the excitatory opsin channelrhodopsin-2 (ChR2), in a semi-closed loop to trigger the light on when the bladder was filled to a level that was lower than the threshold required for spontaneous voiding reflex (light on at 3.1 ± 0.03 min after the previous voiding; inter-voiding interval under control condition: 6.6 ± 0.5 min; Figure 3A; see Methods for detail). This test was also accompanied by a blank control in the absence of ChR2. Light stimulation initiated voiding event at 100% reliability (100%\100%-100%, n = 172 trials of 8 mice; Figure 3B and C) in the test group (response latency, 0.7\0.6-0.8 s; Figure 3D), but almost 0% (0%\0%-1.25%, n = 166 trials of 8 mice) in the blank control group (Figure 3B and C). Accordingly, we also performed simultaneous cystometry with the EUS-EMG experiment under urethane anesthesia (Figure 3E; see Methods for detail). Light stimulation induced a prominent upstroke of bladder pressure in the test group and did not affect bladder pressure at all in the control group (ΔP: 5.6\4.5-8.2 cm H₂O, n = 9 mice in the test group, -0.1\-0.2-0.1 cm H₂O, n = 6 mice in the control group, p = 4e-4, Wilcoxon rank-sum test; Figure 3F-H). Meanwhile, light stimulation triggered the phasic bursting activity of EUS-EMG, at a success rate of 100% in the test group and nearly 0% in the control group (Figure 3F-H). An additional set of control experiments by using regular inter-stimulation intervals (5 s light-on per every 30 s) instead of the threshold-adaptive interval, yielded the same 100% success rate of both bladder pressure and EUS-EMG responses (Figure supplement 7). Thus, photoactivation of PMC^ESR1+ cells resulted in initiating the voiding process through both contracting the bladder and relaxing the sphincter. These sets of ‘loss-of-function’ and ‘gain-of-function’ experiments together demonstrate that PMC^ESR1+ cells perform as a 100% reliable ‘master switch’ for either initiate or suspend voiding, provided that all downstream targets are directly controlled and operational, which will be tested next.

Figure 3.

Transection of either the pudendal or pelvic nerve does not impair PMC^ESR1+ neuronal control of the other

To further confirm whether the effect of PMC^ESR1+ cells on the bladder can occur independently of sphincter relaxation, we designed a new set of simultaneous cystometry and EUS-EMG experiments with PMC^ESR1-ChR2 mice subjected to first a pudendal nerve transection (PDNx, disrupting the pathway from the PMC to the EUS) and then additionally a pelvic nerve transection (PLNx, disrupting the pathway from the PMC to the bladder) in the same mice under urethane anesthesia (Figure 4A; see Methods for detail). Under PDNx condition, PMC^ESR1+ cell photoactivation could not elicit EUS-EMG response at all, but robustly elicited bladder pressure upstroke which was then fully abolished after PLNx (Figure 4B-E). Notably, the photoactivation-induced bladder pressure upstroke under the PDNx condition was nearly the same as that in the pre-transection control condition (Figure 4D), suggesting that the PMC^ESR1+ cell control of the bladder was fully operational even when the pudendal nerve to sphincter was severed.

Figure 4.

Given that reflex signals from bladder afferents can indirectly influence the urethral sphincter activity (Chang et al., 2007), we next investigated whether activation of PMC^ESR1+ cells directly induces the bursting pattern of EUS in the absence of bladder contraction. To this end, the sequential nerve transection test was performed again with another group of PMC^ESR1-ChR2 mice in a different order, i.e., first PLNx to disrupt the bladder nerve and then PDNx to additionally disrupt the urethral sphincter nerve (Figure 5A; see Methods for detail). Interestingly, under the PLNx condition with a filled bladder, PMC^ESR1+ cell photoactivation did not induce a bladder pressure upstroke (Δpressure: 3.2\0.96-4.9 cm H₂O in pre-transection control condition, -1.2\-1.4- -0.8 cm H₂O in PLNx condition, n = 8 mice, p = 7.8e-3, Wilcoxon signed-rank test), but the EUS-EMG activity remained responsive, leading to urine leakage which consistently reduced bladder pressure (Figure 5B-E). Both the inversed bladder response and the EUS-EMG response were then fully abolished after the second transection, PDNx (Figure 5B-E-figure supplement 8). Consistently, under the PLNx condition with an unfilled bladder, the EUS-EMG response was present, and the bladder pressure did not change upon PMC^ESR1+ cell photoactivation (Figure supplement 8). Collectively, these data together indicate that PMC^ESR1+ cell control of the urethra sphincter was operational even when the pelvic nerve to the bladder was severed.

Figure 5.

Lastly, despite that the PMC^ESR1+ cell photoactivation consistently elicited EUS-EMG bursting in 100% of cases, a refined analysis revealed that the key parameters of the sphincter were slightly degraded under PLNx condition (EMG bursting AUC area: 0.21\0.20-0.22 mV*s in pre-transection, 0.17\0.15-0.19 mV*s in PLNx; EMG bursting duration: 4.9\4.8-5.03 s in pre-transection, 4.5\4.1-4.6 s in PLNx, n = 8 mice, p = 7.8e-3; Wilcoxon signed-rank test; Figure 5E), which implies a potential reduction in effective voiding volume (Langdale and Grill, 2016). To further explore this, we conducted experiments on awake, unrestrained mice under the PLNx condition (Figure 6A, see Methods for detail). Compared to the pre-transection control condition (‘baseline’), PMC^ESR1+ cell photoactivation still triggered voiding in 100% of trials (Figure 6B and C). However, in PLNx mice, the voiding events were characterized by significantly smaller urine spots and a longer latency to voiding onset than the control events before transection, a difference not observed in the sham surgery group (Figure 6D and E). These findings could be interpreted as that while the PMC^ESR1+-EUS pathway is sufficient to drive voiding, as previously reported (Keller et al., 2018), however, without an intact bladder reflex pathway involved, an efficient urine flow cannot be performed. Piecing these data of the combined transection-optogenetics experiments together, we reveal a more complete picture that PMC^ESR1+ cells can operate the bladder (through the pelvic nerve) and the sphincter (through the pudendal nerve) independently of each other. Such refined knowledge could not have been obtained from simpler tests using optogenetics alone as shown above (Figures 2 and 3), or from other experiments in the literature (Hou et al., 2016; Keller et al., 2018; Ito et al., 2020).

Figure 6.

PMC^ESR1+ cells coordinate bladder contraction and sphincter relaxation to initiate urination

The above data showing that PMC^ESR1+ cells can operate through both the bladder and urethral sphincter independently of each other implies that there should be two distinct anatomical projections from PMC^ESR1+ cells downstream to innervate these targets. Indeed, our anterograde labeling experiment in ESR1-Cre mice (Figure supplement 9; see Methods for detail) showed that the mGFP-labelled axonal terminals of the PMC^ESR1+ cell population were found in both the DGC and the SPN region of the lumbosacral spinal cord, which further send axons via the pudendal and pelvic nerves to innervate urethral sphincter and bladder, respectively. To further test this hypothesis, we performed retrograde labeling experiments in ESR1-Cre mice and, as a control, in CRH-Cre mice (Figure 7A and E; see Methods for detail). Analysis of CRH-Cre mice showed that 82.4% (80.1%\78.7%-86.7%) of PMC^CRH+ cells were SPN-projecting only (expressing mCherry), and 1.8% (1.4%\1.1%-2.2%) were DGC-projecting only (expressing EGFP), whereas the remaining 15.8% (16.8%\11.4%-19.1%) were dual-projecting (altogether n = 1416 cells pooled from 79 slices obtained from 3 mice; Figure 7B-D). This result is consistent with the literature (Valentino et al., 2010; Ito et al., 2020; Hou et al., 2016; Keller et al., 2018) that PMC^CRH+ cells primarily project to the SPN in the spinal cord and modulate bladder contraction. By contrast, analysis of the ESR1-Cre mice showed that 19.2% (24.1%\13.9%-24.2%) of PMC^ESR1+ cells were SPN-projecting only, 53.3% (50.9%\41.2%-55.2%) were DGC-projecting only, and 27.5% (26.8%\25%-30.9%) cells were dual-projecting (n = 1417 cells pooled from 82 slices obtained from 5 mice; Figure 7F-H). These data suggest that PMC^ESR1+ cells innervate either SPN or DGC in a more balanced manner and include a significant fraction of dual-innervating cells, in contrast to PMC^CRH+ cells that primarily innervate the SPN.

Figure 7.

The functional and anatomical data above together suggest that PMC^ESR1+ cells can function as a 100% reliable ‘master switch’ either to initiate or to suspend voiding through independently operating the bladder (via SPN to the pelvic nerve) and the sphincter (via DGC to the pudendal nerve). We now come to the final question as to whether PMC^ESR1+ cells can implement the coordination between the bladder and urethra sphincter, i.e., operate them in a rigid temporal order. We took advantage of the simultaneous recording of cystometry and EUS-EMG condition (e.g. Figure 2) in which the beginning/ending timepoint of bladder pressure upstroke (a significant rapid, transient increase in bladder pressure preceding voiding, denoting threshold pressure of bladder contraction) (Rana et al., 2024) and the stereotypic voiding-associated firing pattern of EUS-EMG (denoting sphincter relaxation) could be precisely determined. In the photometry recording experiments when voiding events spontaneously occurred, the bladder pressure upstroke timepoint always preceded the beginning timepoint of the EUS-EMG firing pattern (bladder pressure upstroke onset: 0.6\0.3-1.2 s, EMG bursting onset: 2.3\1.5-3.2 s, relative to the onset of Ca²⁺ signals; n = 46 events pooled from 8 mice; Figure 8A). Accordingly, in the photoactivation experiments, the bladder pressure upstroke timepoint also preceded the EMG bursting pattern begin timepoint, albeit both were slightly earlier than those in ‘passive’ spontaneous photometry recordings (bladder pressure upstroke onset: 0.4\0.4-0.5 s, EMG bursting onset: 0.8\0.7-1.0 s; n = 50 events pooled from 10 mice; Figure 8B). Despite that the photoactivation of the PMC^ESR1+ cells could have been artificially strong and did not necessarily mimic the naturalistic firing pattern of PMC^ESR1+ cells (see Figure 1G-I), the same temporal order as bladder contraction preceded sphincter relaxation suggests that the downstream circuity was instructed to execute in the same temporal order. Intriguingly, in photoinhibition experiments, the EMG pattern almost instantaneously ended, while the bladder pressure upstroke end timepoint occurred later (EMG bursting end: 0.1\0.1-0.2 s; bladder pressure upstroke end: 0.8\0.8-0.9 s, n = 44 events pooled from 8 mice; Figure 8C). These timing results can be interpreted from a broader perspective by considering the lower spinal reflex circuit together with PMC^ESR1+ projections from the brainstem (Figure 8D). Since the lower spinal reflex circuit is the forefront agent that can send a feedforward reflex signal from the bladder to the sphincter unidirectionally (Lee et al., 2021; de Groat et al., 2015; Chang et al., 2007; Abud et al., 2015), the brainstem coordination signal descending in two parallel paths from PMC^ESR1+ cells would result in the initiation of voiding with the same forwarding order but in the suspension of ongoing voiding with the reversed order.

Figure 8.

Discussion

A prominent aspect of our data is that voiding events 100% correlated with PMC^ESR1+ neuronal activity (Figure 1), and both photoactivation and photoinhibition of PMC^ESR1+ cells yielded a 100% success rate in initiating and suspending the process of voiding, respectively (Figure 2 and 3). Whilst there are many other factors from both the lower spinal circuit and the higher cortical/subcortical circuits together to determine in what condition to void (Yao et al., 2018; de Groat et al., 2015; Sartori et al., 2022; de Groat, 2009; Zderic, 2019; Mukhopadhyay and Stowers, 2020), the instantaneous execution to efficiently initiate or suspend a voiding process involves a significant contribution of PMC^ESR1+ cells in the brainstem as demonstrated here in this study. The bidirectional control, i.e., in the one way to initiate voiding whenever needed and suitable (which is a basic life need) (Mukhopadhyay and Stowers, 2020), and in the other way to suspend an ongoing voiding when needed (e.g., to release only a small volume of urine for landmarking) (Keller et al., 2018; Hyun et al., 2021; Desjardins et al., 1973), can be executed at 100% reliability by PMC^ESR1+ cells given that all downstream nerves and muscles are intact and operational.

Moreover, the anatomical data also revealed a significant fraction of dual-projecting PMC^ESR1+ cells (Figure 7H) which we consider to play a significant role in the coordination of the bladder and sphincter. Because these two projections originate from the same cells, the ‘on’ and ‘off’ commands for bladder contraction sent from these cells would always align with the ‘on’ and ‘off’ commands for sphincter relaxation, respectively. As such, the brainstem can serve the upstream higher brain regions (Yao et al., 2018; Zare et al., 2019; Manohar et al., 2017) (e.g., the neocortex) with a 100% reliable ‘master switch’ to urinate a desired volume whenever in physiological or social needs (Mukhopadhyay and Stowers, 2020; Malykhina, 2017; Kaur et al., 2014), e.g., to mark territories with small urine spots as an animal, or to urinate into a small sample tube for a health check as a human.

The fact that PMC^ESR1+ cells can execute dynamic, real-time control of voiding at 100% reliability does not imply that other cell types in the PMC are not required to achieve the same goal of perfect urination control. On the contrary, we suggest that a proper baseline control of bladder pressure is also mandatory, e.g., by CRH⁺ cells which are the majority of cells in PMC that primarily operate the bladder through the pelvic nerve and can be modulated by various contextual factors (Hou et al., 2016; Vincent and Satoh, 1984; Wood et al., 2009). After all, it is not the cell type domain marker ESR1 (which by itself abundantly expressed in many other regions of the brain and is involved in many other physiological and cognitive functions (Karigo et al., 2021; Liu et al., 2022; Fang et al., 2018)), but rather the topology of innervation (Figure 8D) that directly enables such a role of urination coordination, i.e., a cell of whichever marker that locates in this particular brainstem nucleus (PMC) (Kawatani et al., 2021) and possesses dual innervations towards both the bladder and sphincter is a potent contributor to urination coordination (Griffiths, 2015). Nevertheless, our pinpointing of PMC^ESR1+ cells performing a 100% reliable role in bladder-sphincter coordination enlightens the future development of precise therapies for treating urination coordination disorders.

Conclusions

In summary, we have identified the essential role of PMC^ESR1 neurons in coordinating urination through co-innervation of both the bladder and urethra sphincter. Our findings offer new insights into the anatomical and physiological basis and research paradigms for the coordinated engagement of parasympathetic and somatic functions of urination control. PMC^ESR1 neurons may serve as a key focal point for advancing our understanding of the neural mechanisms underlying urination, both in physiological contexts and pathological conditions such as brain injuries, spinal cord injuries, and peripheral nerve damages.

Materials and methods

Animals

The experiment procedures were approved by the Third Military Medical University Animal Care and Use Committee and were conducted strictly in adherence to established guidelines. This study utilized ESR1-IRES-Cre (Jackson Laboratory, stock #017911) (Lee et al., 2014) and CRH-IRES-Cre (Jackson Laboratory, stock #012704) (Baram et al., 2015) mice. These mice were group-housed in an environment-controlled room at 23-25℃ and 50% humidity, with 4-5 mice per cage, on a 12-hour light/dark cycle, and with free access to food and water. Mice implanted with optical fibers were housed individually. Both male and female mice, aged 8-20 weeks, were randomly assigned to various experiments. The figure legends and supplement tables detail the number of animals utilized in each experiment.

Virus vectors and CTB

The study utilized the following viruses and cholera toxin subunit B (CTB) for various experiments: For fiber photometry recording experiments, AAV2/9-DIO-GCaMP6f (titer: 0.5 × 10¹² vg/ml) was unilaterally injected into the PMC. For optogenetic manipulation experiments, AAV2/9-hsyn-DIO-hGtACR1-mCherry (titer: 1.43 × 10¹³ vg/ml) and AAV2/8-DIO-ChR2-mCherry (titer: 1.33 × 10¹³ vg/ml) were bilaterally delivered into the PMC for photoinhibition and photoactivation, respectively. For retrograde tracing experiments, pAAV2/retro-EF1a-DIO-EGFP (titer: ≥ 1.00 × 10¹² vg/ml) and pAAV2/retro-EF1a-DIO-mCherry (titer: ≥1.00 × 10¹² vg/ml) were injected into the spinal cords of ESR1-Cre and CRH-Cre mice, respectively. For anterograde tracing experiments, AAV2/9-hEF1a-fDIO-mGFP (titer: 1.00 × 10¹³ vg/ml) was injected unilaterally into the PMC, and a 10: 1 volume mixture of AAV2/2Retro-hsyn-FLEX-Flpo (titer: 1.00 × 10¹³ vg/ml) and CTB 555 (0.2%, C34776, Thermo Fisher; used solely for determining the injection site) was injected into the spinal cord. In control experiments, rAAV2/9-EF1a-DIO-EYFP (titer: 4.20 × 10¹² vg/ml) and rAAV2/9-EF1a-DIO-mCherry (titer: 5.28 × 10¹² vg/ml) were used. All viruses were purchased from Obio Biotechnology Co., Ltd. (Shanghai, China), Taitool Bioscience Co., Ltd. (Shanghai, China), or BrainVTA Co., Ltd. (Wuhan, China).

Stereotaxic injections and optical fiber implant

To target PMC^ESR1+ neurons, ESR1-Cre mice were anesthetized with isoflurane (3% for induction and 1.5–2% for maintenance) and head-fixed in a stereotaxic frame (RWD Life Science Co., Ltd.; Shenzhen, China). Body temperature was maintained at 36°C throughout the surgery using a heating pad. Local anesthesia (lidocaine, 6 mg/kg, subcutaneous injection) was administered at the incision site before making the incision. A small cranial hole above the PMC was created using a dental drill. Approximately 80 nl of the viral solution was delivered to the injection site at a controlled rate of 20 nl/min, either unilaterally or bilaterally, using a micro-syringe pump connected to a glass pipette. The coordinates for PMC injections were: anteroposterior (AP) is -5.45 mm, mediolateral (ML) is ± 0.70 mm, and dorsoventral (DV) is -3.14 mm from the dura. The pipette was then slowly withdrawn over 5 min to prevent virus overflow. The incision was closed with sutures, and the mice received antibiotics and analgesics post-surgery. A heating pad was used for the mice to aid recovery from anesthesia. The mice were then group-housed in their home cages for 3-4 weeks to allow for viral expression.

For optic fiber implantation, the optic fiber (NA: 0.48, diameter: 200 μm, Doric lenses, Quebec City, QC, Canada) was fixed into a metal cannula and positioned 50 μm above the PMC injection site. The fiber was implanted unilaterally for photometry experiments and bilaterally for optogenetics experiments. Dental cement was used to affix the optical fiber to the skull. Mice were housed individually and given 3-5 days to recover before recording or stimulation sessions. After the experiments, the placement of the virus and optical fiber was confirmed by histology in each mouse.

Targeted spinal cord injections

For spinal cord injection surgery, the method previously described was used (Chen et al., 2019). Briefly, mice were anesthetized under isoflurane (1.5-2% oxygen) and positioned on a heating pad. After shaving the hair, a midline skin incision (1-2 cm) was performed over the lumbar segments following local anesthesia (lidocaine, 6 mg/kg, subcutaneous injection) at the incision site. The tissue and muscle connected to the dorsal spine were dissected to expose the T12-L2 vertebrae. The spine was affixed to a stereotactic frame using a spinal adapter (68094, RWD), and the spinal cord between L1 and L2 was exposed by removing the ligamentous and epidural membranes. Using the central vein as a reference, a total of 80 nL of virus solution was injected at two different locations, each injection (40 nL) was administered through a micro-syringe pump connected to a glass pipette, targeting DGC (+ 0.12 mm lateral to the central vein, 0.52 mm deep from the dorsal surface at a 10-degrees angle) and SPNs (- 0.12 mm lateral to the central vein, 0.52 mm deep from the dorsal surface at a 30-degrees angle). The pipette remained in position for a minimum of 5 min before being slowly removed to prevent any leakage. The injection sites were sealed with tissue glue (Vetbond, 3M Animal Care Products), followed by suturing of the skin. Mice received antibiotics and analgesics post-surgery. Approximately 3-4 weeks after the injections, the brain and spinal cords were extracted for histological validation. In ESR1-Cre or CRH-Cre mice, PMC cells projecting to the SPN (labeled with mCherry) and DGC (labeled with EGFP) regions of the spinal cord were manually quantified using Image J.

Pudendal nerve and pelvic nerve transection

Pudendal nerve transection was performed as described previously (Khorramirouz et al., 2016; Peh et al., 2018). Briefly, mice were anesthetized with isoflurane (1.5-2% oxygen) and positioned on a heating pad. A midline skin incision was made along the back from L4 to the coccyx, followed by paraspinal incisions through the gluteal muscles and fascia to expose the sciatic nerve. The sciatic nerve was gently retracted to expose the pudendal nerve. Using microsurgical scissors, the bilateral pudendal nerves, along with the anastomotic branch, were carefully dissected and excised.

For pelvic nerve transection, modifications to the established method (Chang et al., 2018) were made to minimize surgical trauma. Mice were positioned laterally, and bilateral paraspinal incisions were extended upward. The sciatic and pudendal nerves were gently retracted laterally to expose the pelvic nerve, which was identified (originating from the sacral segments of the spinal cord and connecting to the major pelvic ganglion) and severed. For pelvic nerve transection experiments in freely moving mice, the muscles and skin were sutured, and antibiotics and analgesics were administered. In the sham injury group, the same procedures were followed, except for the transection of the pelvic nerve.

Fiber photometry recording and analysis in freely behaving mice

The Ca²⁺ recordings of PMC^ESR1+ neurons were conducted using a fiber photometry setup, as described previously (Yao et al., 2018; Rao et al., 2022). Fluorescence at the fiber tip was excited by blue light (0.22 mW/mm²). Mice with implanted fibers were injected intraperitoneally with diuretics (furosemide, 40 mg/kg) and acclimated in a testing chamber equipped with a bottom camera (1,280 × 720 pixels) for 20 min before recording. Signals from PMC^ESR1+ neurons and voiding behavior were simultaneously recorded for approximately 40 min. Ca²⁺ signals were sampled at 2000 Hz using NI LabVIEW software (National Instruments, USA), while behavioral video was captured at 30 Hz. Fiber photometry data and video were synchronized via event markers. For data analysis, all signals were processed with a Savitzky-Golay filter (third-order polynomial, 50 side points) for low-pass filtering. Δf/f = (f - f_baseline)/f_baseline was calculated to assess photometry signals during voiding, where f_baseline represents the minimum fluorescence recorded. Results were presented as heatmaps using MATLAB. Ca²⁺ signal data were shuffled by dividing the original dataset into 10 segments and randomly associating them with voiding events. Positive signals were defined as Ca²⁺ signal amplitudes exceeding three times the noise band (the standard deviation). This procedure was also applied to control mice to correct for movement artifacts.

Single-unit with optrode recording and analysis

To identify the single-unit activity of PMC^ESR1+ neurons, optrode recordings were performed as described previously (Qin et al., 2022; Qin et al., 2018; Yang et al., 2023). Briefly, the optrode consisted of a 200 µm optical fiber and four tetrode assemblies aligned in a line, spaced 100 µm apart. The optical fiber was secured to the electrodes, positioned 500 µm above their tips, and connected to an LED. Each electrode assembly comprised four twisted tungsten wires (25 µm, California Fine Wire), allowing vertical movement via micromanipulators. Optrode implantation surgery was performed in ESR1-Cre mice expressing ChR2 in the PMC, with the electrode tips aligned and implanted 2.80 mm below the brain dura. After a recovery period of 5-7 days, the tetrodes were slowly inserted to a target depth of -2.95 mm and recording began. Single-unit signals from PMC^ESR1+ neurons in freely behaving mice were recorded using an RHD2000 USB board (C3100, Intan Technology) at 20 kHz, while behavioral video was captured simultaneously. Units with short spike latencies (< 7 ms) in response to light pulses of varying intensities (5 mW, 10 mW, 15 mW, and 20 mW) and high responsiveness (> 70 %) were identified as PMC^ESR1+ cells. To confirm recording locations, electro-lesions were performed by applying a current (10 μA, 12 s) through the tetrodes.

The raw recorded data were preprocessed using established methods to extract peaks (Qin et al., 2018). All events exceeding the amplitude threshold (set at four standard deviations above the background) were kept for further analysis. The average firing rate of each cell was calculated within a sliding time bin of 20 seconds around voiding (0.1 s intervals), divided by the total number of trails, and adjusted by subtracting the baseline value (the median firing rate during the -10 to -5 s before voiding). The results were visualized as a heatmap (logarithmic analysis) in Figure 1. Statistical analysis of the average firing rates over a 2-second interval before voiding (from -10 s to -8 s) and around voiding (from -1 s to 1 s) was conducted for PMC^ESR1+ neurons and non-PMC^ESR1+ neurons, respectively.

Optogenetic experiments in freely behaving mice

Before optogenetic stimulation, mice underwent the same procedures as for fiber photometry: they received a diuretic with the intraperitoneal injection to increase urination events and were acclimated to the testing chamber for at least 20 min. For optogenetic inhibition experiments, bilateral stimulation (1 mW/mm² at fiber tips, 50 Hz frequency, 20 ms pulses) was delivered using 473 nm blue light. To assess the effect of light inhibition on urination, mice were placed in a glass chamber (28 cm × 16 cm × 30 cm) with a 0.19 mm filter paper (14.6 cm × 27 cm, BWD) underneath. Urination was observed at three stages: pre-photoinhibition (light-off), during the 5 s or 60 s of photoinhibition (light-on), and post-photoinhibition (light-off). The photoinhibition parameters (frequency and duration) were controlled via the NI LabVIEW platform (National Instruments, USA). Real-time urination behavior was monitored simultaneously using two cameras positioned above and below the glass chamber (Chongqing NewLight Co., Ltd; see: www.newlightxhr.com). The laser was manually triggered at the onset of urination (within 4 s), and experiments where the trigger exceeded 4 s were excluded. Urination cessation was defined as the initiation of movement after urination. The void area, total void duration, and latency were analyzed from the video and are presented in Figure 2 and Figures supplement 5 and 6.

For optogenetic activation experiments, bilaterally fiber-implanted mice were connected to two 473-nm blue laser generators (5 mW/mm² at fiber tips, 25 Hz frequency, 15 ms pulses) via optic fibers. During testing, blue light was delivered for 5 s with approximately 3-min intervals between trials, over a 30-45 min session. This light stimulation was repeated over two days with a one-day interval. Mouse behavior was monitored and recorded simultaneously. The stimulation was performed before and after transection for photoactivation experiments in freely behaving mice with pelvic nerve transection using the same procedures. Note that the photostimulation experiments for pelvic nerve transactions were conducted one day after the surgery. Detailed success rates, latency of voiding, and void area after photoactivation are shown in Figures 3 and 6.

Simultaneous cystometry and electromyography in anesthetized mice

Bladder catheter and urethral sphincter electrode implantation were performed as previously described (Verstegen et al., 2019; Hou et al., 2016; Keller et al., 2018). Briefly, adult fiber-implanted ESR1-Cre mice were anesthetized with isoflurane and placed on a heating pad. A lower mid-abdomen incision exposed the bladder and urethral sphincter. PE-10 tubing was inserted through the bladder dome and secured with a 6-0 Ethicon suture. For electromyography (EMG) recording, two 160 μm silver-plated copper wire electrodes (P/N B34-1000, USA) with 2 mm exposed tips were inserted into the external urethral sphincter (EUS) on both sides using a 30-gauge needle, positioned between the urethra and pubic symphysis, and spaced at least 2 mm apart. A ground wire with 4 mm exposed tips was inserted subcutaneously near the sternal notch to minimize signal interference. The ends of the bladder catheter tubing and wire electrodes were exteriorized through an incision on the skin at the back of the neck, and both abdominal and neck incisions were closed. For measuring intravesical pressure, the bladder tubing was connected to a pressure transducer (YPJ01H; Chengdu Instrument Factory, China) and a syringe pump (RWD404; RWD Technology Corp., Ltd., China) via three-way stopcocks. Bladder pressure and EUS-EMG data were recorded through a multi-channel physiological recording device (RM6240; Chengdu Instrument Factory, China) sampled at 8 kHz. After surgery, mice were permitted to recover from anesthesia and resume walking.

For simultaneous recording, all fiber-implanted mice were anesthetized with urethane (1.2 g/kg, i.p.) and continuously infused with room-temperature physiological saline at 30-50 µl/min via the bladder catheter for at least 45 min. Recording or stimulation was performed once regular bladder pressure cycles associated with natural urination events were established. “Filling bladder” was defined as continuous saline infusion, while “non-filling bladder” was defined as no infusion. Cystometry and EUS-EMG data were captured via commercial acquisition software, alongside monitoring of Ca²⁺ signals and mouse behavior. For cross-correlation analysis, cystometric data, EMG data, and photometry data were first downsampled to 40 Hz and standardized using z-scored. The original EMG data were processed to extract their envelope using MATLAB’s “envelope” function. The original data were merged, and divided into 10 segments, and segments of photometry data were randomly matched with segments of cystometry or envelope EMG data to create shuffled datasets. Cross-correlations of cystometric and photometry data, or EMG and photometry data were calculated using MATLAB’s “xcorr” function, and the peak values of cross-corrections were reported, as shown in Supplementary Figure 3.

For photoactivation experiments performed simultaneously with cystometry and electromyography recording, blue light pulses were delivered periodically (every 30 s) at 25 Hz for 15 ms, lasting 5 s, or randomly at intervals between 20 s and 40 s, with each condition repeated at least 15 times. For photoactivation experiments performed simultaneously with cystometry and urethral electromyography recording under pelvic or pudendal nerve transection conditions, the surgical procedures and pre-recording preparations were the same as those described above, with the following changes: Mice underwent three stages of randomized photostimulation (consisted of 25 Hz, 15 ms, 5-s durations, with intervals between 30 s and 60 s) in sequence under urethane anesthesia (1.2 g/kg, i.p.): an intact nerve period, a pudendal nerve or pelvic nerve transection period, and a period with both pudendal nerve and pelvic nerve transection. All trials were pooled to assess the impact of nerve transection on bladder and sphincter function for each mouse. Cystometric data, including ΔP = P_{5 sec} -P_{0 sec} (where P_{0 sec} is the pressure at the onset of laser stimulation and P_{5 sec} is the pressure at the end of laser stimulation) and the pressure ratio P_ratio = P_{max (0-5)sec} / P_{mean (-5-0)sec} (where P_{max (0-5)sec} is the maximum pressure from laser onset to cessation, and P_{mean (-5-0)sec} is the average pressure during the 5 seconds preceding laser onset), were calculated using MATLAB. The EUS-EMG data (burst duration and area under the curve) were analyzed using multi-channel physiological recording software. The spectrogram was generated using envelope EMG data around photostimulation in MATLAB, as shown in Figures 4 and 5, and Figure supplement 8. The onset time of bladder pressure upstroke was identified by finding the maximum of the second derivative of cystometry curves around pressure peaks using a MATLAB script. The onset time of EMG bursting (Kadekawa et al., 2016b; Cheng, 2004) was manually defined as the points at which bursting activity begins. The results are shown in Figure 8.

For photoinhibition experiments performed simultaneously with cystometry and electromyography recording, the light was manually triggered at the onset of EMG bursts and delivered in 20 ms pulses at 50 Hz for either 5 s or 60 s. Cystometric data, including Δpressure = P_peak – P_min (where P_peak is the peak pressure and P_min is the minimum pressure after the peak) and threshold pressure (bladder pressure upstroke), were analyzed manually using the multi-channel physiological recording software. The EUS-EMG data (burst duration, area under the curve, and latency of termination) was analyzed using the multi-channel physiological recording software. The latency of bursting termination is defined as the interval from laser onset to the end of EUS-EMG bursting. The end time of bladder pressure upstroke (termination of the rapid increase in bladder pressure before the end of voiding, denoting bladder relaxation) was identified by finding the minimum of the second derivative of cystometry curves around pressure peaks using a MATLAB script. The end time of EMG bursting was manually defined as the points at which the onset of tonic activity after bursting. The results are shown in Figures 3 and 8.

Histology and immunohistochemistry

Following the completion of all experiments, histological verification of fiber implantation or virus injection positions was conducted. Mice were deeply anesthetized with 1% sodium pentobarbital (10 ml/kg), followed by transcardial perfusion with cold 0.9% saline and 4% paraformaldehyde (PFA). The brain or spinal cord (in some experiments) was then extracted and post-fixed overnight at 4 °C in ice-cold 4% PFA. Coronal brain sections (40 µm) and the thoracolumbar and lumbosacral segments of spinal cord sections (70 µm) were cut using a freezing microtome. For TH immunohistochemistry, free-floating sections were initially incubated in 1% PBST (1% Triton X-100 in PBS, Sigma) for 60 min. The sections were blocked with 10% donkey serum (Sigma) in 0.1% PBST for 2 hours at room temperature. Following blocking, then incubated at 4°C for 24 hours with anti-TH antibodies (1:200 dilution, rabbit, Sigma-Aldrich). After extensive washing with PBS, sections were incubated with a secondary antibody (1:300, Alexa Fluor 488 or 594 donkey anti-rabbit, Invitrogen) at room temperature for 2 hours. Finally, all sections, including the target segments of the spinal cord, were incubated with DAPI (1:1000, Beyotime) for 15 min. Images were acquired using a confocal microscope (TCS SP5, Leica) equipped with × 4, × 10, and × 20 oil objectives, utilizing 405 nm, 488 nm, and 552 nm lasers, or with an Olympus microscope.

Statistical analysis

All data were processed and statistically analyzed using Prism 8 GraphPad, MATLAB, and SPSS 22 software. For unpaired group comparisons, the Wilcoxon rank-sum test was used, and for paired groups, the Wilcoxon signed-rank test was applied, as described in the figure legends. Analysis was performed by investigators blinded to the experiments. The n value reflects the final number of animals in each experiment group, with animals excluded if histological verification of gene expression showed poor or absent. Data are represented as median with 25%-75% percentiles, and statistical significance was defined as *< 0.05, **< 0.01, ***< 0.001; ns, no significant difference.

Acknowledgements

The authors are grateful to Ms. Jia Lou for her help in composing and editing the layout of the figures. This study was supported by grants from the National Natural Science Foundation of China to X.C. (No. 31925018, 32127801), the National Key R&D Program of China to X.C. (2021YFA0805000), the Suzhou Science and Technology Plan Project (SZS2022008), and the Jiangsu Provincial Big Science Facility Initiative (BM2022010), and the Guangxi Talent Program (“Highland of Innovation Talents”). X.C. is a member of CAS Center for Excellence in Brain Science and Intelligence Technology.

Additional information

Funding

Author contributions

Project design, J.W.Y, and X.C.; injection and histology, X.L., X.P.L., J.L.(1), L.X.Y., and J.L.(2); behavior experiments, X.L., C.H.Y., and X.W.; nerves transection, X.L.; fiber recording, X.L. C.H.Y., and X.P.L.; electrophysiology, X.L., H.Q., T.L.J., and X.W.; fiber recording, X.L. and L.X.Y.; cystometry and electromyography experiments, X.L., X.P.L., and J.L.(1); data interpretation and analysis, X.L., H.Q., S.S.L., H.B.J., X.Liao, J.W.Y., and X.C.; figure preparation, X.L., H.B.J., J.W.Y, and X.C.; manuscript writing, X.L., H.B.J., J.W.Y, and X.C. with the help of all co-authors. All authors read and commented on the manuscript.

Ethics

All experiment procedures were approved by the Third Military Medical University Animal Care and Use Committee (Approval number: AMUWEC20230061) and were conducted strictly in adherence to established guidelines.

Data availability

Any additional information underlying the findings of this study is available from the corresponding authors upon reasonable request. The supporting data underlying Figures 1-8 and Figures supplement 1-9 are provided as Source Data files. Source Data are provided in this paper.

Code availability

The codes supporting the current study have not been deposited in a public repository, but are available from the corresponding author upon request.

Figures supplement

Figure supplement 1.

Figure supplement 2.

Figure supplement 3.

Figure supplement 4.

Figure supplement 5.

Figure supplement 6.

Figure supplement 7.

Figure supplement

Figure supplement 9.