Acetazolamide-Induced Brain Acidosis as a Candidate Dual-Target Neuroplastic Intervention: A Theoretical Framework for Recalibrating Respiratory Chemoreception and ASIC1a-Mediated Amygdala Fear Circuitry

Locke

This document proposes that time-limited acetazolamide therapy may simultaneously recalibrate two pH-sensitive neural systems co-dysregulated by chronic stress: the retrotrapezoid nucleus (respiratory control) and the amygdala (fear/anxiety). Both are sensitized by the same upstream driver (orexin), both detect brain pH through specific acid-sensing mechanisms, and both may undergo durable neuroplastic adaptation under sustained mild acidosis. Published evidence supports many of the component links independently, while the integrated dual-recalibration mechanism remains untested. This document connects them into a unified, testable hypothesis intended for expert evaluation, not clinical implementation.

Two Co-Occurring Conditions

Chronic sleep fragmentation and chronic anxiety frequently co-occur in patients under sustained psychological stress. The conventional interpretation treats them as comorbid but mechanistically independent: poor sleep worsens anxiety, anxiety worsens sleep. This document proposes that a subset of such patients may share a single upstream mechanism operating through two pH-sensitive neural circuits, and that a single pharmacological intervention could recalibrate both simultaneously.

The Unified Disease Model

Orexin as the Upstream Driver

Chronic psychological stress activates orexin (hypocretin) neurons in the lateral hypothalamus. Orexin is a neuropeptide released in response to stress, emotional arousal, and cognitive stimulation. Chronic psychological stress and cognitive/emotional arousal persistently elevate orexinergic tone. Orexin does not broadcast diffusely; it projects to specific downstream targets with specific consequences. Two of those targets are central to this proposal.

Target 1: The Retrotrapezoid Nucleus (Respiratory Control)

The retrotrapezoid nucleus (RTN) is a specialized cluster of approximately 5,000 neurons (bilaterally in humans; Kumar et al., 2025) located at the ventral surface of the medulla oblongata, ventral to the facial motor nucleus. RTN neurons are identified by expression of the transcription factor Phox2b, vesicular glutamate transporter 2 (VGLUT2), and the absence of tyrosine hydroxylase and choline acetyltransferase. They are the brain's primary central respiratory chemoreceptors: they detect changes in brain extracellular fluid (ECF) pH and drive the ventilatory response by sending excitatory glutamatergic signals to the ventral respiratory column (VRC), which contains the respiratory pattern generators (the pre-Botzinger and Botzinger complexes).

RTN neurons detect pH through two specific proton-sensing mechanisms. TASK-2 channels (TWIK-related acid-sensitive K⁺ channel 2) are background potassium leak channels that are directly inhibited by extracellular H⁺. When pH drops, TASK-2 channels close, trapping K⁺ intracellularly, depolarizing the neuron, and increasing its firing rate. TASK-2 knockout severely blunts central respiratory drive; the channel handles pH sensitivity in approximately 44% of Phox2b⁺ cells. GPR4 (G protein-coupled receptor 4) is a proton-activated G_s-coupled receptor. GPR4 deletion severely disrupts RTN neuron activation, increases spontaneous apneas, and blunts the ventilatory response to CO₂. CRISPR/Cas9-engineered mutations of the specific histidine residues responsible for pH sensing on GPR4 completely eliminated pH sensitivity in vitro and CO₂-stimulated breathing in vivo, confirming that GPR4 is fundamentally a proton sensor, not a CO₂-specific sensor. In isolated brainstem preparations, isocapnic metabolic acidosis (lowered bicarbonate at constant PCO₂) activates RTN neurons comparably to hypercapnic acidosis. The RTN detects H⁺ regardless of its physiological source.

Orexin binds directly to receptors on RTN neurons, increasing their sensitivity to CO₂/H⁺ (Lazarenko et al., 2011). Li and Nattie (2010) confirmed that the dual orexin receptor antagonist almorexant attenuates the CO₂ ventilatory response by 26% during wakefulness. When this orexin-mediated sensitization operates chronically, the subject overbreathes not by conscious choice but because the chemoreceptors are being pharmacologically driven into treating normal CO₂ levels as insufficient. The increased ventilation is not matched by increased metabolic CO₂ production. The breathing is clearing CO₂ that the muscles are not generating.

Chronic hyperventilation depletes arterial CO₂. Over days to weeks, the kidneys respond to the sustained hypocapnia by excreting bicarbonate to normalize blood pH, and the choroid plexus simultaneously reduces bicarbonate concentration in the cerebrospinal fluid. Because the CSF lacks the robust protein buffers found in systemic blood, the loss of bicarbonate strips the chemoreceptors of their chemical shock absorbers. Tiny CO₂ fluctuations that a normally buffered system would absorb without consequence now produce massive, unbuffered pH swings in the fluid surrounding the central chemoreceptors. The chemoreceptors recalibrate to the depleted CO₂ environment as the new normal. Controller gain rises. The CO₂ reserve, the gap between resting PaCO₂ and the apneic threshold, narrows to almost nothing.

During sleep, when cortical override is absent and respiration falls entirely under autonomic chemical control, this miscalibrated system produces ventilatory instability. Normal NREM CO₂ fluctuations that a healthy system would absorb trigger microarousals, sleep-stage shifts, central pauses, and architectural fragmentation. Sleep quality is destroyed despite adequate CPAP treatment for any obstructive component. Standard monitoring (AHI, pulse oximetry) may show nothing abnormal because the dominant events are sub-threshold CO₂/H⁺ micro-instabilities that fragment sleep without producing scoreable apneas or oxygen desaturation. This is the high-controller-gain sleep fragmentation phenotype.

Target 2: The Amygdala (Fear and Anxiety)

Orexin neurons project densely to the amygdala and the bed nucleus of the stria terminalis (BNST), promoting hyperarousal, panic, and anxiety (Fagan et al., 2024; Han et al., 2019). Independently of its orexin inputs, the amygdala senses brain pH through acid-sensing ion channel 1a (ASIC1a).

Ziemann et al. (2009), published in Cell, demonstrated that CO₂ inhalation reduces brain pH and activates ASIC1a in the amygdala, producing intense fear and avoidance behavior. This response requires functional ASIC1a; genetic knockout abolishes it. Viral-mediated overexpression of ASIC1a in the amygdala of null mice fully rescues the CO₂-evoked fear response. The BNST is equally critical: BNST lesions or localized ASIC1a disruptions attenuate CO₂-evoked panic (Taugher et al., 2014). Post-acquisition CO₂ inhalation actively enhances cued and contextual fear memory 24 hours later, an effect entirely dependent on ASIC1a (Taugher et al., 2021). The amygdala is therefore not merely responding to fear stimuli processed elsewhere; it is itself a chemosensor that directly detects brain pH and generates fear behavior in response to acidosis.

In a system already primed by chronic orexin exposure, the amygdala's ASIC1a-mediated pH sensitivity is layered on top of orexin-driven hyperactivation. Normal CO₂/pH fluctuations that would be physiologically unremarkable in a non-sensitized system may be sufficient to trigger ASIC1a-mediated fear and anxiety responses. This is a model prediction, not an established finding.

An important nuance: ASIC1a is expressed on both glutamatergic principal neurons and GABAergic interneurons within the basolateral amygdala. The net behavioral output depends on the balance between direct pyramidal excitation and ASIC1a-driven feedforward inhibition (Pidoplichko et al., 2014). Acute, intense pH drops may drive massive pyramidal cell excitation and fear. Sustained, lower-level acidosis may preferentially recruit inhibitory interneurons, exerting an anxiolytic braking effect. This principal-versus-interneuron balance is directly relevant to why acetazolamide's gradual acidosis may produce anxiolysis rather than fear: the sustained, mild stimulus may bias the network toward inhibition rather than excitation. However, the converse risk must be acknowledged: if ASIC1a desensitizes more effectively on GABAergic interneurons than on excitatory pyramidal cells, the net effect could be amygdala disinhibition, producing increased rather than decreased fear reactivity. The direction of this balance under sustained mild acidosis is not established and must be determined empirically.

The Self-Reinforcing Loop

The two targets create a closed, self-reinforcing loop. Orexin-driven RTN sensitization produces sleep fragmentation. Sleep fragmentation impairs REM-dependent emotional processing: during REM, norepinephrine levels drop to their lowest point, allowing emotional memories to be reprocessed without the accompanying stress response. Without adequate REM, emotional material is stored with its full affective charge intact, returning the next day as raw, unintegrated reactivity. Unprocessed emotional material sustains threat sensitivity, rumination, and anxiety. The anxiety sustains orexin output. Orexin continues to sensitize both the RTN and the amygdala. The respiratory and psychiatric symptoms are not comorbid by coincidence; they share an upstream driver and reinforce each other through a closed loop.

The Clinical Intersection: Published Evidence for Linked CO₂ Sensitivity

The co-occurrence of sleep-disordered breathing and anxiety/panic disorders is well established epidemiologically but typically treated as comorbid rather than mechanistically linked. Several lines of published evidence suggest a deeper connection through shared CO₂/pH sensitivity.

First, CO₂ hypersensitivity is a documented feature of both conditions. Elevated loop gain (the respiratory control system's tendency to overshoot) characterizes high-controller-gain sleep fragmentation. Exaggerated fear responses to CO₂ inhalation characterize panic disorder. Both reflect a system that overreacts to normal CO₂/pH fluctuations, one in the respiratory domain, the other in the fear domain. Messineo et al. (2018) validated that reduced voluntary breath-hold duration (the Control Pause) independently predicts elevated loop gain during NREM sleep, establishing a behavioral link between CO₂ intolerance and respiratory instability.

Second, CO₂ habituation studies show that repeated CO₂ exposure produces simultaneous respiratory and anxiety adaptation. Beck et al. (1999) found that 37% of high-anxiety subjects habituated their fear and autonomic responses across repeated CO₂ challenges. If the habituation involved adaptation of pH-sensitive circuits in both the RTN (respiratory) and amygdala (fear), the dual benefit is mechanistically coherent under the model proposed here.

Third, orexin has been implicated in both conditions independently. Orexin drives chemoreceptor sensitization in the RTN (Lazarenko et al., 2011; Li and Nattie, 2010) and fear amplification in the amygdala (Fagan et al., 2024; Han et al., 2019). The shared upstream driver provides a mechanistic basis for co-occurrence that goes beyond statistical comorbidity.

Acetazolamide: Sustained Brain Acidosis Uncoupled from Arterial Blood Gases

What Acetazolamide Does

Acetazolamide is a carbonic anhydrase (CA) inhibitor, FDA-approved for decades for altitude sickness, glaucoma, epilepsy, and central sleep apnea. It blocks bicarbonate reabsorption in the renal tubules, producing a sustained mild metabolic acidosis. It also crosses the blood-brain barrier and inhibits CA locally within the brain parenchyma, choroid plexus, and cerebral vasculature.

The Ventilatory Compensation Problem and Its Resolution

A critical objection: acetazolamide-induced metabolic acidosis stimulates ventilation. The patient hyperventilates. Arterial PaCO₂ drops. Because CO₂ freely diffuses across the blood-brain barrier, dropping arterial PaCO₂ should draw CO₂ out of the brain and alkalize it. If this happens, the RTN and amygdala never experience sustained acidosis, and the entire neuroplasticity hypothesis fails.

Bickler et al. (1988) resolved this directly. In mechanically ventilated subjects with compensatory hyperventilation simulated, they measured brain extracellular fluid pH simultaneously with arterial blood gases. The results were definitive. Arterial PaCO₂ dropped by 8.0 +/- 5.9 Torr from hyperventilation. Brain ECF PCO₂ returned to essentially baseline (delta -0.2 +/- 1.6 Torr). Yet brain ECF pH still fell by 0.08 +/- 0.04 pH units. Brain ECF bicarbonate dropped by 1.7 +/- 1.3 mM.

The explanation: local CA inhibition in the brain traps H⁺ that cannot be cleared through the normal CO₂ dehydration pathway. The brain's ability to convert H⁺ and bicarbonate back into CO₂ (which could then diffuse out) is blocked by the enzyme inhibition. H⁺ accumulates locally. Brain ECF pH and arterial PaCO₂ are fundamentally uncoupled. The brain remains acidified despite compensatory reductions in arterial PaCO₂.

An important caveat: Bickler measured acute brain pH changes. Over a multi-week treatment course, the brain's homeostatic mechanisms, including choroid plexus acid-base transport and astrocytic Na⁺/H⁺

exchangers, will actively attempt to normalize brain ECF pH. Whether the acetazolamide-induced acidosis is sustained at therapeutically meaningful levels through the full treatment period, or whether glial buffering partially neutralizes the effect within days, is an open question that must be resolved by in vivo pH monitoring across the full treatment duration in the proposed animal study.

The uncoupled brain acidosis potentially affects pH-sensitive neural populations across multiple brain regions simultaneously: the RTN, the amygdala, the BNST, and any other circuit with proton-sensing machinery. The drug produces a sustained, brain-wide pH perturbation that persists for as long as it is administered.

The Respiratory Side: RTN Galanin-Mediated Recalibration

Galanin in the RTN

Approximately 50% of RTN Phox2b⁺ neurons co-express the neuropeptide galanin (Stornetta et al., 2008). Galanin is a 29-30 amino acid inhibitory neuromodulator. When released from RTN axon terminals, it acts on Galanin Receptor 1 (GalR1), a G_i/G_o-coupled receptor expressed on neurons of the ventral respiratory column: approximately 30% of glycinergic neurons in the Botzinger complex and 11% in the pre-Botzinger complex (Dereli et al., 2019). GalR1 activation inhibits adenylyl cyclase, lowers intracellular cAMP, opens GIRK channels, and inhibits calcium channels, producing potent, long-lasting inhibition of neuronal excitability. Pharmacological microinjection of galanin directly into the pre-Botzinger complex drastically attenuates ventilation, reduces the CO₂-mediated chemoreflex, and can induce total apnea. Galanin receptor antagonists (galantide, M40) reverse this inhibition. Non-peptide synthetic galanin agonists (galnon, galmic) mimic the ventilatory suppression.

Biphasic Response Under Chronic Acidosis

Dereli et al. (2019) demonstrated that exposure to chronic hypercapnia produces a biphasic neuropeptide response in the RTN. Short-term hypercapnia (6-8 hours) decreases preprogalanin mRNA by approximately 33%, removing the inhibitory brake to permit an acute hyperventilatory escape response. Long-term hypercapnia (10 days at 5-8% CO₂) produces the opposite: a 40-60% increase in preprogalanin mRNA expression. Critically, the galanin-expressing RTN neurons maintained their full responsiveness to acute CO₂ challenge after chronic exposure, while numerous other respiratory-related neuronal populations showed blunted activation. The RTN galanin system adapts without losing its core chemosensory function.

The proposed mechanism: chronically stimulated RTN neurons upregulate galanin production. Galanin is co-released with glutamate from the same synaptic terminals. The galanin binds GalR1 on downstream VRC pattern generators and actively suppresses the excitatory signal. Same CO₂ input, less ventilatory output. Controller gain drops at the molecular level. Galanin functions as a homeostatic governor that prevents runaway hyperventilation.

Translation to Humans

A landmark 2025 bioRxiv preprint by Kumar, Shepherd, Machaalani, and colleagues characterized the adult human RTN for the first time. Using immunohistochemistry on formalin-fixed human brainstem tissue, they identified approximately 5,000 Phox2b⁺/TH^-/ChAT^- neurons bilaterally, located ventral to the facial motor nucleus, matching the canonical rodent RTN definition. 43% of these human RTN neurons expressed galanin, closely mirroring the 45-50% rate in rodents and the approximately 2,400 galaninergic RTN neurons mapped in macaques. The anatomical substrate and molecular machinery required for acetazolamide-induced galaninergic neuroplasticity are present in the adult human brain at quantitatively comparable levels.

Clinical Durability Evidence

Verbraecken et al. (1998) studied eight patients with non-hypercapnic central sleep apnea who received one month of low-dose acetazolamide (250 mg/day). The Central Apnea Index (CAI, the number of central apnea events per hour of sleep) was measured at baseline, after one month of treatment, after one month off drug, and after six months off drug. Baseline CAI: 25 +/- 10. After treatment: 4 +/- 2. After one month off drug: 5 +/- 3. After six months off drug: 3 +/- 1.

Central apneas dropped by 84% during treatment and remained suppressed for six months after discontinuation, actually improving slightly between the one-month and six-month off-drug measurements.

While promising, this N=8 uncontrolled study cannot exclude regression to the mean or delayed physiologic adaptation; the signal is striking but requires replication. The authors explicitly hypothesized a long-lasting resetting of the central chemoreceptor drive. The study also noted a morphological shift from pure central apneas to obstructive events, suggesting that while upper airway anatomy remained unchanged, the CNS propensity to trigger total apneic pauses had been fundamentally altered.

DeBacker et al. (1995) reported similar findings in 14 patients with idiopathic central sleep apnea: AHI dropped from 37.2 to 12.8 on treatment, with benefits sustained at one-month off-drug follow-up. White et al. (1982) demonstrated acute improvement with higher doses (1000 mg/day) over shorter duration, though long-term durability tracking was less extensive.

Acetazolamide's plasma half-life is measured in hours. Renal bicarbonate typically recovers within 1-2 weeks in patients with healthy kidneys. There is no pharmacokinetic or acid-base explanation for six-month persistence. The galanin upregulation hypothesis provides a specific, transcriptionally grounded candidate mechanism: structural gene expression changes in RTN neurons that outlast the drug and its metabolic effects.

The COPD Negative Control

Wagenaar et al. (1996) investigated acetazolamide in hypercapnic COPD patients. Unlike the Verbraecken findings, all respiratory parameters returned to baseline one month after discontinuation. The authors could not confirm persistent chemoreceptor resetting. The galanin hypothesis predicts this divergence: in COPD patients, the RTN is already chronically exposed to high PaCO₂ and localized acidosis. The neuroplastic capacity for galanin upregulation may be saturated or exhausted. In non-hypercapnic CSA patients, the RTN baseline is normocapnic or mildly hypocapnic; acetazolamide-induced acidosis acts as a novel stimulus triggering de novo adaptation. Durability is phenotype-specific.

Competing Explanations for Durability

The Verbraecken durability could theoretically reflect mechanisms other than galanin upregulation: slowly normalizing bicarbonate dynamics (insufficient to explain six months), structural synaptic remodeling in the RTN (astrocytic changes, dendritic spine remodeling), or recalibration of the NMB/galanin excitatory-inhibitory neuropeptide balance rather than galanin alone. The galanin hypothesis is a leading candidate because it provides a specific, anatomically localized, transcriptionally grounded mechanism with published molecular evidence at every link. It is not, however, proven. The critical first test is the animal experiment proposed in section “The Testable Experiment”.

Acute Stabilization Versus Durable Recalibration

These are distinct hypotheses and must be evaluated separately. Acute stabilization (improvement while taking acetazolamide) can be fully explained by CO₂-reserve widening through metabolic acidosis. This is established and not novel. Durable recalibration (persistent improvement after discontinuation, beyond the timeframe of acid-base normalization) requires a structural explanation. The galanin hypothesis provides one. The central endpoint is not improvement while on the drug; it is persistence after discontinuation, ideally accompanied by reduced hypercapnic ventilatory response, increased Control Pause, reduced loop gain, and preserved sleep architecture off medication.

The Amygdala Side: ASIC1a-Mediated Fear Circuit Recalibration

ASIC1a as the Amygdala's pH Sensor

Acid-sensing ion channel 1a (ASIC1a) is a proton-gated cation channel abundantly expressed in the amygdala, particularly the basolateral nucleus, and in the BNST. Ziemann et al. (2009), in a study published in Cell and cited over 550 times, established that the amygdala is itself a chemosensor: CO₂ inhalation reduces brain pH, activates ASIC1a in the amygdala, and produces fear behavior. This is not a secondary response to respiratory distress; it is a direct pH-to-fear transduction pathway. Knockout of ASIC1a abolishes the response. Viral overexpression in the amygdala of null mice rescues it.

The BNST and Fear Memory Consolidation

Taugher et al. (2014) demonstrated that the BNST, an extra-amygdalar structure rich in ASIC1a, is critical for anxiety-related behavior evoked by CO₂; BNST lesions or localized ASIC1a disruptions attenuate CO₂-evoked panic. Taugher et al. (2021) showed that post-acquisition CO₂ inhalation actively enhances cued and contextual fear memory 24 hours later, and this enhancement is entirely ASIC1a-dependent. ASIC1a therefore participates not only in acute fear responses but in the consolidation of fear memories, making it relevant to conditions like PTSD where fear memories are pathologically persistent.

Fear Memory Lability and Extinction

Recent studies have revealed that ASIC1a participates not only in fear acquisition but in memory lability and extinction. Transient acidosis induced by CO₂ inhalation during retrieval of a fear memory makes the memory trace highly labile through an ASIC1a-dependent mechanism. This lability involves insertion of calcium-permeable AMPA receptors (CP-AMPARs), increased CREB phosphorylation, and reactivation of a greater percentage of neurons in the memory trace. The labile state renders the memory susceptible to modification or enhanced extinction. If acetazolamide maintains a sustained, mild acidotic environment, fear memories accessed during the treatment period may exist in a chronically labile state, potentially facilitating their extinction or reconsolidation in a less threatening form. However, fear-memory lability is not inherently therapeutic. A reactivated labile memory can be weakened, modified, or strengthened depending on context, arousal, and subsequent experience. This suggests that therapeutic benefit from acetazolamide-induced lability may require structured behavioral context (extinction training, safety learning) rather than drug-alone administration. Without such context, increased lability could theoretically permit maladaptive reconsolidation. This is directly relevant to PTSD, where pathologically persistent fear memories resist extinction.

Steady-State Desensitization Under Chronic Acidosis

ASIC1a exhibits fundamentally different behavior under acute versus chronic pH challenges. An acute drop in pH triggers rapid channel opening, ion flux, and downstream signaling. However, under continuous or sustained acidosis, or pre-exposure to mild pH drops (such as pH 7.2 to 7.0), ASIC1a undergoes steady-state desensitization (SSD): the channel transitions directly from a closed state to a non-conducting desensitized state without ever opening (Chu and Xiong, 2012; Duan et al., 2011). Chronic low-level acidosis functionally silences ASIC1a without triggering the acute fear response. Endogenous modulators such as spermine and dynorphins can alter the pH dependence of SSD, potentially modulating the susceptibility of different individuals to this adaptation.

This is the critical mechanistic distinction for the amygdala hypothesis. Acetazolamide produces gradual, sustained brain acidosis, not the abrupt pH drop of CO₂ inhalation. SSD predicts that the amygdala's ASIC1a channels would progressively desensitize under this chronic stimulus rather than chronically firing.

The fear circuit quiets rather than activates.

Tachyphylaxis: A Deeper Desensitized State

Beyond standard SSD, ASIC1a exhibits tachyphylaxis: a progressive, long-lasting decay in peak current amplitude across repeated acid stimulations, even when the inter-stimulus interval allows theoretical recovery from standard desensitization. Tachyphylaxis represents a distinct, slow desensitized state that is intrinsically linked to channel opening. Any transient openings of ASIC1a triggered by synaptic activity during sustained mild acidosis will progressively drive the receptor pool into deep tachyphylaxis, rendering the amygdala functionally unreactive to acute CO₂ or stress-induced proton spikes. The magnitude of tachyphylaxis increases with deeper acidic conditioning, increased stimulation frequency, and prolonged acid exposure.

ASIC1a Internalization and Dendritic Spine Remodeling

Prolonged desensitization is followed by physical receptor removal. Under chronic acidosis, ASIC1a undergoes internalization through a clathrin- and dynamin-dependent endocytic pathway. The N-terminus of the channel alters its binding conformation, unbinding from the C-terminus within residues 463-476, facilitating recruitment of intracellular trafficking machinery that pulls the channel out of the synaptic membrane. This serves an evolutionary neuroprotective function: prolonged ASIC1a activation mediates Ca²⁺ overload and acidosis-induced excitotoxicity, so the neuron actively removes the channel to survive sustained metabolic stress.

ASIC1a also regulates structural remodeling of dendritic spines. Reduced surface trafficking protects against acidosis-induced spine loss. A time-limited course of acetazolamide therapy is therefore likely to trigger not only functional desensitization (SSD and tachyphylaxis) but physical receptor internalization and morphological synaptic remodeling in the amygdala. Reversing these changes requires new protein synthesis, membrane trafficking, and spine reconstruction, processes that operate on timescales of weeks to months. This provides a biophysical basis for durability after drug discontinuation.

The Three-Layer Desensitization Cascade

SSD, tachyphylaxis, and internalization form a three-layer cascade, each deeper and more durable than the last. Layer 1 (SSD): channels transition to a non-conducting state under sustained mild acidosis without opening; reversible within hours if pH normalizes. Layer 2 (tachyphylaxis): channels that do open under synaptic activity enter a long-lived desensitized state; recovery takes days. Layer 3 (internalization): channels are physically removed from the membrane through endocytic trafficking; recovery requires new protein synthesis and trafficking, potentially weeks to months. Acetazolamide's sustained brain acidosis would engage all three layers simultaneously, producing a progressive, durable reduction in ASIC1a-mediated fear reactivity that deepens over the treatment period and persists well beyond drug discontinuation.

Convergent Evidence: IV Acetazolamide Does Not Trigger Panic

Consistent with the SSD prediction, intravenous acetazolamide does not trigger panic attacks in panic disorder patients despite producing brain acidosis. This was demonstrated in double-blind trials by both Mathew et al. (1989) and Gorman et al. (1993). The Gorman study noted that patients actually showed a trend toward decreased anxiety and dyspnea following acetazolamide administration. Meanwhile, CO₂ inhalation, which produces abrupt pH drops, reliably triggers panic in the same population. The gradual versus abrupt distinction in the pH stimulus maps directly onto the SSD versus acute-activation distinction in ASIC1a physiology. The drug acidifies the brain without triggering the fear circuit because the acidosis is slow enough to drive desensitization rather than activation.

Bicarbonate Depletion as an Additional Mechanism

Acetazolamide depletes brain ECF bicarbonate (Bickler measured a 1.7 mM drop). Bicarbonate modulates ASIC1a function: administering bicarbonate prevents ASIC1a-mediated pathology in traumatic brain injury models (Yin et al., 2013), suggesting bicarbonate serves a protective role. Acetazolamide-induced bicarbonate depletion may alter ASIC1a gating and pH exposure, but the net direction of this effect is uncertain. It could deepen desensitization under controlled mild acidosis, but it could also reduce protection against acid stress. This should be treated as a mechanistic variable to measure in the proposed experiment, not assumed as beneficial.

Direct Evidence: Acetazolamide Already Affects Amygdala Fear Circuitry

Yang et al. (2013) demonstrated that acetazolamide impairs fear memory consolidation in rodents. Animals given acetazolamide showed significantly decreased avoidance learning, corroborated by reduced fear conditioning-induced long-term potentiation (LTP) in the amygdala. The authors explicitly proposed application to PTSD. This is published direct evidence that acetazolamide affects amygdala fear processing at the synaptic plasticity level.

An important distinction: Yang showed impairment of fear memory consolidation, the process of forming new fear memories. Treating established anxiety and PTSD requires fear extinction, the process of learning that a previously threatening stimulus is now safe. These are neurobiologically distinct processes involving different molecular cascades. The consolidation impairment Yang demonstrated does not directly establish that acetazolamide facilitates extinction of existing fears. However, the ASIC1a SSD mechanism proposed here (sections “Steady-State Desensitization Under Chronic Acidosis”-6.7) and the fear memory lability effect (section “Fear Memory Lability and Extinction”) operate through different pathways than consolidation impairment and may independently facilitate extinction by reducing ASIC1a-mediated threat signaling and rendering existing fear traces more susceptible to modification.

What Yang did not address: whether the effect is durable after discontinuation, whether ASIC1a steady-state desensitization is the mechanism, whether a time-limited course could produce persistent recalibration, or whether the drug facilitates extinction of existing fears rather than merely impairing consolidation of new ones. Those are the questions this proposal asks.

A Second Amygdala Proton Sensor: GPR68

The proton-sensitive GPCR GPR68 is also expressed in the amygdala and contributes to fear-related behavior and hippocampal LTP. Its expression regulates transcriptome changes in stress-response and neuroprotection pathways during acidosis (Zhou et al., 2021). Acetazolamide's brain acidosis would affect GPR68-expressing amygdala neurons alongside ASIC1a-expressing ones, potentially engaging multiple acid-sensing pathways simultaneously.

CO₂ Habituation and Its Failures

Repeated CO₂ inhalation produces robust habituation of fear and autonomic responses in approximately 37% of high-anxiety individuals (Beck et al., 1999), functioning as interoceptive desensitization. However, panic disorder patients often demonstrate an inability to habituate tidal volume across repeated CO₂ exposures (Blechert et al., 2010). The SSD model suggests a mechanistic explanation: successful habituators may achieve ASIC1a desensitization under repeated exposure, while non-habituators may have ASIC1a gating properties (possibly modulated by endogenous factors like spermine or dynorphins) that resist SSD. Acetazolamide could theoretically force the biological adaptation that non-habituating panic patients cannot achieve through exposure alone, by providing the sustained, gradual acidotic stimulus that favors SSD over acute activation.

Additional suggestive clinical observations, not part of the core mechanistic chain but consistent with the broader hypothesis: low-dose acetazolamide has resolved treatment-resistant premenstrual dysphoric disorder with severe comorbid anxiety (Sani et al.). In bipolar disorder, a scoping review found a 44% response rate among patients with severe mania or refractory depression, with responders maintaining benefit for up to two years, which is consistent with neuroplastic adaptation but could reflect other mechanisms. These observations are heterogeneous and should not be treated as direct evidence for the ASIC1a hypothesis.

One Drug, Two Recalibrations

The respiratory and psychiatric symptoms are downstream manifestations of the same upstream dysregulation. A drug that acidifies the brain addresses both because both depend on pH-sensitive circuits that were miscalibrated by the same chronic orexin exposure. This is not a coincidental dual indication. It is a unified mechanism with two clinical faces.

An important asymmetry must be acknowledged: the respiratory and amygdala halves of this hypothesis are not equally established. The respiratory side has stronger clinical durability evidence through the Verbraecken acetazolamide study. The amygdala side has strong channel-level mechanistic plausibility and direct acetazolamide/fear-memory evidence (Yang 2013), but lacks comparable longitudinal clinical durability data. The dual-recalibration model should be treated as asymmetric: respiratory recalibration is the more directly supported arm; amygdala recalibration is the higher-risk, higher-novelty extension.

The Orexin Layer: DORA as Complementary Agent

In the unified disease model, orexin sensitizes both the RTN and the amygdala. Even if acetazolamide produces adaptive changes in both systems, ongoing orexin bombardment may continuously re-sensitize them, fighting the adaptation. Dual orexin receptor antagonists (DORAs) block this signal.

A DORA may serve three complementary roles: (1) acute reduction of orexin-driven hyperactivation in both the RTN and the amygdala; (2) sleep protection by reducing arousal frequency and consolidating sleep architecture during the recalibration period, which itself supports emotional processing through restored REM access; (3) permissive plasticity by removing the orexin-driven re-sensitization signal that maintains both pathological states. The third role is speculative.

Suvorexant reduces anxiety- and panic-like behaviors in animal models and improves mood in psychiatric inpatients (Han et al., 2019). For daytime coverage of the orexin-driven sensitization window, the agent must have a half-life that provides meaningful daytime receptor occupancy. Lemborexant (half-life 17-19 hours) is the only approved DORA that meets this requirement. Taken at bedtime, it consolidates sleep while providing residual daytime coverage sufficient to blunt stress-induced orexin peaks without causing frank narcolepsy.

A potential pharmacokinetic interaction requires evaluation. DORAs (suvorexant, daridorexant) are metabolized by the hepatic enzyme CYP3A4. Acetazolamide has been reported to inhibit CYP3A4, which could increase DORA serum concentration and CNS presence. Whether this interaction is clinically meaningful at therapeutic acetazolamide doses is not established and must be verified through pharmacokinetic studies before any combination protocol. If confirmed, it represents a safety variable requiring dose adjustment and monitoring, and it could potentially allow lower DORA doses to achieve adequate receptor occupancy. It should not be assumed as a strategic advantage without verification.

The DORA is proposed as an enhancement to acetazolamide monotherapy, not a co-equal partner. Acetazolamide provides the primary recalibrating stimulus through sustained brain acidosis. The DORA creates a more permissive environment for that recalibration to take hold by removing the force that maintains the pathological state.

The Testable Experiment

A single mouse cohort can test both hypotheses simultaneously at minimal additional cost beyond a respiratory-only study. Acetazolamide for 10 days at doses producing mild metabolic acidosis. Environmental hypercapnia positive control. Vehicle control. All measurements repeated post-washout to assess persistence.

RTN harvest: Preprogalanin mRNA via qPCR. GalR1 receptor density in VRC (to assess potential downstream receptor downregulation from chronic galanin exposure). CO₂ ventilatory response before, during, and after washout. Arterial/CSF pH, PaCO₂, bicarbonate.

Amygdala harvest (basolateral amygdala): ASIC1a mRNA and protein levels. ASIC1a surface expression versus internalized pools (to assess desensitization-driven receptor trafficking). Galanin, neuropeptide Y, and other inhibitory neuropeptide expression. GPR68 expression. CO₂-evoked fear behavior on-drug and after washout if feasible. Amygdala LTP if feasible.

Pharmacological and genetic controls: PcTx1 (psalmotoxin), a highly specific ASIC1a inhibitor, as a positive control for ASIC1a-dependent effects. Amiloride and its analogs (benzamil, phenamil) for concentration-dependent ASIC1a blockade (Lingueglia and Lazdunski, 2013; Leng et al., 2016). If feasible, an acetazolamide + ASIC1a knockout or ASIC1a blockade arm would distinguish direct ASIC1a-mediated effects from other acetazolamide mechanisms. Multiple washout timepoints (24 hours, 7 days, 30 days) would characterize the durability profile of any observed changes.

Behavioral controls: Open Field Test and/or Rotarod to confirm normal locomotion and rule out sickness behavior from systemic acidosis masquerading as reduced fear. A mouse showing reduced CO₂-evoked freezing must also demonstrate normal exploratory behavior and motor function to establish that the effect is specific to fear circuitry, not generalized malaise.

In vivo pH monitoring: Optical or microdialysis pH sensors in brain ECF across the full 10-day treatment period to verify that acetazolamide maintains sustained acidosis and that glial buffering mechanisms do not neutralize the effect by mid-course.

Fear extinction protocol: Beyond basic CO₂-evoked fear behavior, the study should include a full fear conditioning followed by extinction paradigm while on the drug, to test whether acetazolamide facilitates extinction of existing fear memories, not merely impairs consolidation of new ones. This distinction is critical for translation to PTSD.

Key endpoint: Do adaptive changes in either system persist after drug discontinuation? Two brain regions from one experiment, two potential publications, one shared mechanistic framework.

Staged Research Design

Stage 0 (Retrospective epidemiological validation): Acetazolamide is prescribed chronically for idiopathic intracranial hypertension (IIH) and other conditions. Electronic health record databases (e.g., TriNetX) contain thousands of such patients. A retrospective analysis could determine whether patients who received extended acetazolamide courses show statistically significant reductions in coded anxiety encounters, sleep disorder diagnoses, or anxiolytic/SSRI prescriptions in the 6-12 months following their drug course compared to matched controls. This costs nothing, requires no new patients, and could provide immediate large-scale epidemiological signal before any animal is dosed.

Stage 1 (Animal, combined RTN + amygdala): Single mouse cohort. Acetazolamide 10 days, low dose to minimize risk of downstream receptor downregulation. Harvest RTN for galanin qPCR and amygdala for ASIC1a expression. Include hypercapnia control, vehicle control, post-washout timepoint. Two publications from one experiment.

Stage 2 (Human, respiratory monotherapy): Acetazolamide in phenotyped high-loop-gain sleep fragmentation patients (CPAP-treated, low AHI, low Control Pause, normal oxygenation, no COPD or baseline hypercapnia). Low dose (125-250 mg/day) using extended-release formulation (e.g., Diamox Sequels) to maintain continuous, smooth brain pH reduction without the peaks and valleys of immediate-release dosing, which could trigger acute ASIC1a activation rather than steady-state desensitization. Multi-week taper at discontinuation to prevent transient plant-gain rebound from bicarbonate depletion. Pre/on/post-discontinuation PSG with transcutaneous CO₂ and daytime end-tidal CO₂. Key endpoint: persistence after washout.

Stage 3 (Human, psychiatric endpoints): Acetazolamide in patients with treatment-resistant anxiety, panic disorder, or PTSD. Given that acetazolamide-induced fear memory lability could theoretically facilitate either extinction or maladaptive reconsolidation depending on context, human psychiatric trials should pair the drug with structured exposure therapy or cognitive behavioral therapy (CBT) rather than administering it as monotherapy. CO₂-evoked fear responses, generalized anxiety scales, panic frequency before, during, and after a time-limited course. Key endpoint: persistence after washout.

Stage 4 (DORA add-on): Lemborexant added in both respiratory and psychiatric cohorts after monotherapy signals are interpretable. Compare monotherapy versus combination for speed, magnitude, and durability of response.

Why This Matters

Existing pharmacological treatments for anxiety, panic disorder, and PTSD (SSRIs, benzodiazepines, buspirone) primarily manage symptoms during ongoing use. Some approaches, including exposure-based psychotherapy and emerging paradigms (ketamine, MDMA-assisted therapy), do target neuroplasticity, but none operate through the specific mechanism of sustained brain acidosis driving adaptive desensitization of pH-sensitive fear circuits. If acetazolamide produces durable ASIC1a-mediated desensitization of amygdala fear circuits, it represents a fundamentally different class of psychiatric intervention: a neuroplastic recalibrator rather than a symptom manager.

The drug is off-patent, cheap, FDA-approved, and has decades of safety data. The course would be weeks, not indefinite. The mechanism is grounded in published acid-sensing neurobiology. This reframes CO₂/pH neuroscience from a niche basic-science field into the mechanistic foundation for a novel approach to treatment-resistant anxiety. The ASIC1a work establishing the amygdala as a chemosensor for CO₂ and acidosis becomes the molecular basis for a pharmacological intervention that nobody has proposed.

Evidence Classification for the Amygdala Hypothesis

Established or strongly supported: ASIC1a mediates CO₂/acidosis fear responses in the amygdala and BNST (Ziemann 2009, Taugher 2014). ASIC1a undergoes steady-state desensitization, tachyphylaxis, and internalization under chronic acid conditions (Chu/Xiong 2012, Duan 2011). Acetazolamide impairs fear memory consolidation and amygdala LTP acutely (Yang 2013). IV acetazolamide does not trigger panic despite brain acidosis (Mathew 1989, Gorman 1993).

Plausible but not established: Acetazolamide at clinical doses produces sufficient amygdala pH shift in vivo to drive therapeutic ASIC1a SSD. SSD/tachyphylaxis/internalization from acetazolamide would reduce pathological anxiety in behaving animals or humans. The effect would persist after drug discontinuation. The effect would spare adaptive fear learning.

Speculative: Chronic acetazolamide could force the biological adaptation that non-habituating panic patients cannot achieve through CO₂ exposure alone. Fear memories may exist in a therapeutically useful chronically labile state during treatment. DORA plus acetazolamide creates a synergistic dual-circuit recalibration window.

Caveats

13.1 ASIC1a steady-state desensitization has been demonstrated in vitro and in cellular preparations. Whether acetazolamide-induced brain acidosis produces sufficient and sustained pH change to drive SSD in vivo in amygdala neurons is the core hypothesis, not an established fact.

13.2 Yang et al. showed acetazolamide impairs fear memory acutely. Whether this represents early ASIC1a desensitization or a separate mechanism is untested. Whether the effect is durable after discontinuation is unknown.

13.3 The respiratory side has Verbraecken's six-month durability data. The amygdala side has no equivalent longitudinal evidence. Durability of ASIC1a adaptation after drug discontinuation is entirely uncharacterized.

13.4 The galanin-amygdala connection has indirect support. Galanin is expressed in amygdala circuits and is anxiolytic (Bajo et al., 2011); galanin expression is dysregulated in PTSD models (Barnabas et al., 2016); and in peripheral neuropathic pain models, ASIC1a downregulation under prolonged tissue acidosis occurs concurrently with galanin and NPY upregulation. GALR2 and NPY Y1 receptor activation in the amygdala specifically reduces freezing. If this co-regulation occurs centrally, acetazolamide-induced acidosis may simultaneously downregulate the excitatory sensor (ASIC1a) and upregulate the inhibitory brake (galanin/NPY). However, direct evidence of this co-regulation in the central amygdala under acetazolamide is absent.

13.5 Chronic ASIC1a desensitization could theoretically impair adaptive fear responses that serve protective functions. The therapeutic goal is recalibration of pathological hyperreactivity, not elimination of normal fear processing.

13.6 The co-occurrence of sleep fragmentation and anxiety under chronic stress is well documented but the proposed shared pH-sensitivity mechanism is a model, not an established finding. Alternative explanations for comorbidity (e.g., HPA axis dysregulation, inflammatory pathways, cognitive-behavioral reinforcement) exist and are not excluded by this framework.

13.7 The acetazolamide + DORA combination for any psychiatric indication has no published precedent. The combination hypothesis is entirely novel and untested.

13.8 The desired therapeutic effect follows an inverted-U curve. Too little brain acidosis may produce no meaningful ASIC1a adaptation. Too much may trigger acute fear activation, cognitive impairment, excessive respiratory drive, or maladaptive plasticity. The therapeutic window is mild, sustained acidosis sufficient to engage adaptive desensitization without crossing into pathological territory.

13.9 The proposed CYP3A4 pharmacokinetic interaction between acetazolamide and DORAs requires formal pharmacokinetic verification. It should be treated as a safety variable, not an assumed advantage.

13.10 The Verbraecken study (N=8, uncontrolled) is the strongest clinical durability evidence and is load-bearing for the respiratory side of the argument. Full-text verification and replication are essential before treating the durability signal as established.

References

Original bibliography wording is preserved. Linked source records identify verified bibliographic corrections; searches are labeled as searches.

1. Bajo M, Madamba SG, Lu X, et al. (2011). Receptor subtype-dependent galanin actions on GABAergic neurotransmission in the central amygdala. Addict Biol, 17(4):694-705.
   Read sourceAbstractOpen full textSource record & notes
   Bibliographic note: The source is a 2012 issue article (online 2011) with a longer title mentioning ethanol responses; the supplied short title is incomplete.
2. Barnabas K, Zhang L, Wang H, et al. (2016). Changes in galanin systems in a rat model of PTSD. PLoS ONE, 11(12):e0167569.
   Read sourceAbstractOpen full textSource record & notes
3. Beck JG, Shipherd JC, Read J. (1999). Response patterns to repeated CO2 inhalation in high anxiety sensitivity. Behav Res Ther, 37(11):1073-1089.
   Read sourceAbstractSource record & notes
4. Bickler PE, Litt L, Banville DL, Severinghaus JW. (1988). Effects of acetazolamide on cerebral acid-base balance. J Appl Physiol, 65(1):422-427.
   Read sourceAbstractSource record & notes
5. Blechert J, Wilhelm FH, Meuret AE, et al. (2010). Respiratory and autonomic responses to repeated CO2 inhalation in panic disorder. Biol Psychol, 84(1):104-111.
   Read sourceAbstractOpen full textSource record & notes
   Bibliographic note: The full verified title includes experiential responses, 20% CO2, social phobia, and healthy controls.
6. Chu XP, Xiong ZG. (2012). Acid-Sensing Ion Channels in Pathological Conditions. Adv Exp Med Biol, 419-431.
   Read sourceAbstractOpen full textSource record & notes
7. DeBacker WA, Verbraecken J, Willemen M, et al. (1995). Central apnea index decreases after prolonged treatment with acetazolamide. Am J Respir Crit Care Med, 151(1):87-91.
   Read sourceAbstractSource record & notes
8. Dereli AS, Yaseen Z, Carrive P, Kumar NN. (2019). Adaptation of Respiratory-Related Brain Regions to Long-Term Hypercapnia. Front Neurosci, 13:1343.
   Read sourceAbstractOpen full textSource record & notes
9. Duan B, Wang YZ, Yang T, et al. (2011). Extracellular spermine exacerbates ischemic neuronal injury through ASIC1a sensitization. J Neurosci, 31(6):2101-2112.
   Read sourceAbstractOpen full textSource record & notes
10. Fagan HA, Huneke NTM, Domschke K, Baldwin DS. (2024). The role of the orexin system in anxiety disorders. Neurosci Applied, 3:103922.
    Read sourceAbstractOpen full textSource record & notes
11. Han Y, Yuan K, Zheng Y, Lu L. (2019). Orexin receptor antagonists as emerging psychiatric treatments. Neurosci Bull, 36(4):432-448.
    Read sourceAbstractOpen full textSource record & notes
    Bibliographic note: The source is a 2020 issue article (online 2019); title is Orexin Receptor Antagonists as Emerging Treatments for Psychiatric Disorders.
12. Kumar NN, Shepherd GM, Machaalani R, et al. (2025). Characterisation of the adult human RTN. bioRxiv preprint.
    Read sourceAbstractOpen full textSource record & notes
    Bibliographic note: The cited preprint shorthand resolves to a peer-reviewed 2025 article with Liu as first author. The published anatomical characterization identifies putative human RTN neurons; it is not a functional intervention study.
13. Lazarenko RM, Stornetta RL, Bayliss DA, Guyenet PG. (2011). Orexin A activates retrotrapezoid neurons. Respir Physiol Neurobiol, 175(2):283-287.
    Read sourceAbstractOpen full textSource record & notes
14. Leng TD, Si HF, Li J, et al. (2016). Amiloride analogs as ASIC1a inhibitors. CNS Neurosci Ther, 22(6):468-476.
    Read sourceAbstractOpen full textSource record & notes
15. Li A, Nattie E. (2010). Almorexant attenuates central chemoreception in wakefulness. J Physiol, 588(Pt 15):2935-2944.
    Read sourceAbstractOpen full textSource record & notes
16. Lingueglia E, Lazdunski M. (2013). Pharmacology of ASIC channels. Wiley Interdiscip Rev Membr Transp Signal, 2(4):155-171.
    Read sourceSource record & notes
17. Messineo L, Taranto-Montemurro L, Sands SA, et al. (2018). Breath-holding as a means to estimate loop gain. J Physiol, 596(17):4043-4056.
    Read sourceAbstractOpen full textSource record & notes
    Bibliographic note: Pharmacological recalibration reference gives a different title and incorrect volume/pages; verified article is 596(17):4043–4056. Maximal breath-hold testing was evaluated as a marker of loop gain, not as a therapeutic training intervention.
18. Stornetta RL, Spirovski D, Moreira TS, et al. (2008). Galanin is a selective marker of the RTN. J Comp Neurol, 512(3):373-383.
    Read sourceAbstractOpen full textSource record & notes
    Bibliographic note: The source is J Comp Neurol 512(3):373–383 (2009; online 2008). Amygdala bibliography uses 2008.
19. Taugher RJ, Lu Y, Wang Y, et al. (2014). BNST is critical for anxiety-related behavior evoked by CO2. J Neurosci, 34(31):10247-10255.
    Read sourceAbstractOpen full textSource record & notes
20. Taugher RJ, Wunsch AM, Wang GZ, et al. (2021). Post-acquisition CO2 enhances fear memory via ASIC1A. Front Behav Neurosci, 15:767426.
    Read sourceAbstractOpen full textSource record & notes
21. Verbraecken J, Willemen M, De Cock W, et al. (1998). Central sleep apnea after interrupting longterm acetazolamide therapy. Respir Physiol, 112(1):59-70.
    Read sourceAbstractSource record & notes
22. Wagenaar M, Vos P, Heijdra YF, et al. (1996). Combined treatment with acetazolamide and medroxyprogesterone in COPD. Eur Respir J, 9:131.
    Search PubMed by titleSource record & notes
    Bibliographic note: Original citation not exactly verified. Related verified source is linked separately; the main link remains a title search.
23. White DP, Zwillich CW, Pickett CK, et al. (1982). Central sleep apnea: improvement with acetazolamide. Arch Intern Med, 142(10):1816-1819.
    Read sourceAbstractSource record & notes
24. Yang MT, Chien WL, Lu DH, Liou HC, Fu WM. (2013). Acetazolamide impairs fear memory consolidation. Neuropharmacology, 67:412-418.
    Read sourceAbstractSource record & notes
25. Yin T, Lindley TE, Albert GW, et al. (2013). Loss of ASIC1a and bicarbonate administration attenuate TBI. PLoS ONE, 8(8):e72379.
    Read sourceAbstractOpen full textSource record & notes
26. Gorman JM, Papp LA, Martinez J, et al. (1993). The effect of acetazolamide on ventilation in panic disorder patients. Am J Psychiatry, 150(10):1480-1484.
    Read sourceAbstractSource record & notes
27. Pidoplichko VI, Bhatt DK, Bhatt DK, et al. (2014). ASICs modulate GABAergic and glutamatergic synaptic transmission in the basolateral amygdala. Neuropharmacology, 77:46-57.
    Search PubMed by titleSource record & notes
    Bibliographic note: Original citation not exactly verified. Related verified source is linked separately; the main link remains a title search.
28. Ziemann AE, Allen JE, Dahdaleh NS, et al. (2009). The amygdala is a chemosensor that detects CO2 and acidosis to elicit fear. Cell, 139(5):1012-1021.
    Read sourceAbstractOpen full textSource record & notes
29. Zhou G, Wang T, Zha X. (2021). RNA-Seq analysis of GPR68 on ischemia-induced transcriptome changes. FASEB J, 35(4).
    Read sourceAbstractOpen full textSource record & notes