Abstract
Rationale
The rostral ventrolateral medulla oblongata (RVLM) is an important region in the brainstem for the neural control of the sympathetic tone and blood pressure. The neurons within the RVLM necessary for these functions remain unidentified, though one population—the catecholaminergic C1 neurons—has been proposed. C1 neurons are activated by lowering blood pressure, they are inhibited by stimulation of the baroreflex, they directly innervate sympathetic preganglionic neurons, and, when artificially stimulated, C1 cells increase sympathetic nerve activity and raise blood pressure. Despite these promising data, recent studies indicate that C1 cells are not required for generating sympathetic tone and maintaining resting blood pressure in conscious animals. As both sympathetic nerve activity and resting blood pressure are thought to be dependent on input from neurons within the RVLM, I explored in my first study whether a second, non-catecholaminergic RVLM group could play a role in generating sympathetic nerve activity and maintaining resting blood pressure.
C1 neurons are no longer thought to be necessary for resting blood pressure maintenance, therefore a different view of C1 in cardiovascular control has developed. Previous studies have indicated that C1 neurons may play an important role in coping with acute stress. C1 neurons are activated during physiological and psychological stress, and inhibition of C1 during physiological stress—but not during quiescence—lowers blood pressure. Further, stimulation of C1 increases sympathetic nerve activity and stimulates release of catecholamines in peripheral circulation, a trademark of the acute stress response (also known as the ‘fight-or-flight’ response). Peripheral catecholamine release promotes survival during acute stress by increasing cardiac output, decreasing blood flow to the digestive system, and increasing blood flow to skeletal muscle. Equally important to coping with an acute stress is the release of catecholamines within the brain, which promotes arousal and increases vigilance. Neuroanatomical data suggests that C1 neurons may innervate central catecholaminergic neurons and activate catecholamine release in the brain, potentially orchestrating a system-wide catecholaminergic response during acute stress. Therefore, in my second study I explored whether C1 neurons directly innervate catecholaminergic neurons in the central nervous system, and what the nature of this innervation could be.
Besides cardiovascular control, the RVLM also plays an important role in the neural control of breathing. Specifically, chemosensitive neurons in the retrotrapezoid nucleus (RTN) regulate blood pH by adjusting breathing in a carbon dioxide-dependent manner. Some data indicate that RTN may also be important to the neural regulation of breathing in general, i.e. independent of the chemosensitive response. The circuits and mechanisms by which RTN control breathing are poorly understood, and while RTN neurons are demonstrably glutamatergic, they also express markers for several signaling molecules that could participate in neurotransmission. Therefore, in my third project I investigated whether glutamate release from RTN was required in regulation of breathing via the RTN.
Methods
In each study, I used the light-gated channelrhodopsin actuator to stimulate a specific group of neurons which were targeted with the Cre-lox system and transgenic mice or viral promoters. Across my three projects, I optogenetically stimulated putatively excitatory neurons of the RVLM, C1 neurons specifically, and a combination of RTN and C1 neurons. I targeted transgene expression to putatively excitatory RVLM by injecting AAV which encodes for the CaMKIIα promoter—which has been claimed to lead to selective expression in excitatory neurons. To specifically target C1 neurons, I used the dopamine-β-hydroxylase Cre+/0 mouse strain (DβH-Cre), which results in Cre expression in only C1 neurons within the RVLM. To target RTN and C1 neurons, I used a lentivirus driven by the artificial promoter PRSx8—which specifically targets C1 and RTN neurons when injected locally into the RVLM.
To measure the consequences of optogenetic stimulation of the RVLM, I recorded aortic pressure in conscious, freely behaving mice by inserting a catheter through the left carotid artery which was attached to a telemetric blood pressure probe. I measured C1-stimulation evoked postsynaptic activity of catecholaminergic neurons by performing whole cell patch clamp recordings in brain slices from reporter mice where catecholaminergic neurons expressed a red florescent protein, and C1 neurons expressed channelrhodopsin-YFP via AAV injection. To measure breathing frequency and volume during stimulation of C1/RTN, I placed mice in a whole body plethysmography chamber—a device which measures changes in airflow, from which breathing data can be extracted.
Results
In DβH-Cre mice, local injection of Cre-dependent caspase AAV resulted in a thorough and specific lesion of C1 neurons. Optogenetic stimulation of the RVLM in conscious DβH-Cre mice robustly increased blood pressure, in both C1-lesioned and C1 intact mice. Under anesthesia, activation of only the spinal axons increased blood pressure, again independent of C1 lesion. In WT mice, injection of retrograde Cre AAV into the spinal cord and injection of Cre-dependent channelrhodopsin AAV into the RVLM resulting in expression of the opsin in bulbospinal neurons—including C1 neurons. Stimulation of these bulbospinal non-C1 and C1 neurons increased blood pressure in conscious mice, while stimulation of C1 specifically in DβH-Cre did not increase blood pressure.
In brain slices, optogenetic stimulation of C1 axons and terminals reliably evoked excitatory postsynaptic currents (EPSCs) and increased firing rates of voltage- and current-clamped catecholaminergic neurons, respectively. Evoked EPSCs were blocked after application of the fast-Na+ channel blocker tetrodotoxin (TTX), which prevents generation of action potentials and action potential-dependent synaptic release. After application of both TTX and 4-aminopyridine (4-AP), optogenetic stimulation was again capable of evoking ESPCs. Finally, application of the glutamatergic blockers cyanquixaline (CNQX) and R-2-amino-5-phosphonopentanoate (AP5) prevented evoked ESPCs by optogenetic stimulation of C1 terminals.
DβH-Cre+/0_Vglut2flox/0 mice bred with Vglut2flox/flox mice generated a conditional knockout cross strain of DβH-Cre+/0_Vglut2flox/flox mice (henceforth, DβH-cKO mice). DβH-cKO mice could not produce detectable Vglut2 protein in DβH neurons via immunohistochemistry, while immunoreactive-Vglut2 protein was detected in both adjacent neurons, as well as in DβH neurons from non-DβH-cKO mice. Selective optogenetic stimulation of C1 in DβH-Cre mice increased breathing, while stimulation of C1 in DβH-cKO did not increase breathing. We then injected PRSx8-channelrhodopsin lentivirus into the RVLM of DβH-Cre and DβH-cKO mice, which resulted in expression of the opsin in both C1 and RTN neurons. Optogenetic stimulation of either mouse group robustly increased breathing, much greater than stimulation of C1 alone. Finally, we injected PRSx8-Cre lentivirus and Cre-dependent channelrhodopsin AAV into the RVLM of either WT or Vglut2flox/flox mice. This resulted in the expression of the opsin in both C1 and RTN, and neurons expressing the opsin in Vglutflox/flox mice lacked detectable Vglut2 protein while those from WT mice had detectable Vglut2 protein. Optogenetic stimulation of C1 and RTN in WT mice robustly increased breathing, while stimulation from Vglut2flox/flox mice produced no change in breathing.
Conclusions
Non-C1 spinally projecting neurons are capable of regulating blood pressure
Optogenetic stimulation of neurons within the RVLM before and after virtually complete lesion of C1 causes a robust increase in blood pressure in conscious mice, demonstrating that non-catecholaminergic RVLM is capable of regulating blood pressure. However, this increase in blood pressure could be due to any number of pathways from RVLM. Under anesthesia, optogenetic stimulation of only the axons in the spinal cord from neurons in the RVLM again increased blood pressure regardless of C1 lesion, demonstrating that selective stimulation only the spinally projecting non-catecholaminergic neurons is sufficient to increase blood pressure, at least under anesthesia. Finally, we stimulated spinally projecting RVLM neurons—both C1 and non-C1—in conscious mice. This again resulted in an increase in blood pressure. We then compared the result of C1 stimulation alone versus stimulation of both C1 and non-C1 bulbospinal neurons. Stimulation of C1 alone did not increase blood pressure, while stimulation of both populations increased blood pressure. Therefore, we conclude that non-catecholaminergic spinally projecting neurons are capable of regulating blood pressure in conscious mice.
C1 monosynaptically activates central catecholaminergic neurons by releasing glutamate
I recorded neuronal activity from central catecholaminergic neurons which were either identified as catecholaminergic via post-hoc IHC or by florescence in DβH-Cre mice crossed with a reporter strain. Once I identified neurons that displayed neuronal activity due to optogenetic stimulation of C1 axons, I tested whether or not this activity was monosynaptic by using TTX. As TTX blocks action potential generation, all polysynaptic activity is abated, and any remaining evoked activity after its application must result from monosynaptic connections. While TTX blocked optogenetically evoked EPSCs in catecholaminergic neurons from C1, application of TTX + 4-AP was sufficient to recover the response while still preventing action potential generation and polysynaptic activity. The addition of 4-AP blocks K+ repolarization, which allows for the optogenetic stimulation of the channelrhodopsin protein to generate enough depolarization to activate voltage-dependent Ca++ channels and therefore induce release of synaptic vesicles. This monosynaptic activation resulted in evoked EPSCs, which were then shown to have been the result of glutamate as blockade of glutamatergic receptors eliminated the evoked EPSCs from optogenetic stimulation of C1 terminals. Therefore, we conclude that C1 monosynaptically activates central catecholaminergic neurons by releasing glutamate.
RTN neurons require Vglut2 to drive breathing
We have previously shown that optogenetic stimulation of C1 neurons in mice in which C1 neurons lack Vglut2 (DβH-Cre+/-_Vglut2flox/flox) does not increase breathing. We wished to know if the increase in breathing due to stimulation of RTN was also dependent on Vglut2, however there is currently no way to specifically target RTN neurons. The artificial PRSx8 promoter specifically targets C1 and RTN neurons when injected locally into the RVLM. Optogenetic stimulation of RTN and C1 after injection of the PRSx8 lentivirus robustly increased breathing in control mice, but also in mice where C1 neurons lack Vglut2 and cannot influence breathing, demonstrating that RTN stimulation is driving breathing in these mice. Then, in WT mice, we stimulated RTN and C1 using a cocktail of PRSx8-Cre lentivirus and Cre-dependent channelrhodopsin, and optogenetic stimulation again produced a large increase in breathing. We injected the same viral cocktail to label RTN and C1 in Vglut2flox/flox mice, which prevents expression of Vglut2 in both C1 and RTN neurons. Optogenetic stimulation in this case produced very little activation of breathing despite many C1 and RTN neurons expressing the channelrhopsin protein. This demonstrates that the increase in breathing due to optogenetic stimulation of RTN requires Vglut2 and is likely dependent on glutamate release.
