Astrocytic gamma-aminobutyric acid dysregulation as a therapeutic target for posttraumatic stress disorder

Abstract

Post-traumatic stress disorder (PTSD) remains a debilitating psychiatric condition with limited pharmacological treatment options. Identifying novel therapeutic targets is critical for addressing its unmet clinical needs. Through our comprehensive human clinical research, including both cross-sectional and longitudinal studies, we revealed a compelling link between dysregulated prefrontal gamma-aminobutyric acid (GABA) levels and PTSD symptoms. Notably, elevated prefrontal GABA levels in PTSD patients are associated with impaired cerebral blood flow (CBF) and symptom severity, normalizing with recovery, highlighting GABA dysregulation as a key mechanism in the disorder. Postmortem and PTSD-like mouse models implicated monoamine oxidase B (MAOB)-dependent astrocytic GABA as a primary driver of this imbalance, exacerbating deficit in fear extinction retrieval. Genetic and pharmacological inhibition of MAOB effectively restored astrocytic GABA and improved fear extinction retrieval in PTSD-like mouse models. Specifically, KDS2010, a recently developed highly selective and reversible MAOB inhibitor, not only restored astrocytic GABA homeostasis but also rescued CBF deficits and reduced tonic GABA and astrogliosis in the prefrontal cortex. Moreover, KDS2010 successfully advanced through Phase 1 clinical trials, showing a favorable safety profile and paving the way for Phase 2 trials to evaluate its therapeutic potential in PTSD. Our findings highlight the pivotal role of astrocytic GABA in PTSD pathophysiology and establish MAOB inhibition as a mechanistically targeted approach to alleviate symptoms. By bridging human and animal studies with translational clinical trials, this work positions KDS2010 as a promising first-in-class therapy, offering a novel paradigm for PTSD treatment.

Fig. 1

Overview of study design and aims. a The first part of this research consisted of two comprehensive human clinical studies aimed at examining alterations in prefrontal GABA levels and their clinical implications in PTSD. Cross-sectional clinical study analyzed a cohort of 248 participants, divided into groups of persistent PTSD (N = 78), recovered PTSD (N = 84), and healthy control (N = 86), to investigate prefrontal GABA alterations in PTSD and their impact on CBF and clinical symptoms. Longitudinal clinical study involved 126 participants, including a trauma-exposed group (N = 65) and a healthy control group (N = 61), to study the reversibility of prefrontal GABA alterations as PTSD symptoms improved over time. b The second part of this research explored the therapeutic potential and underlying cellular mechanisms of interventions targeting prefrontal GABA inhibition for PTSD treatment. Postmortem human brain study focused on elucidating the cellular mechanisms underlying prefrontal GABA alterations in PTSD through the analysis of postmortem human brain tissues. PTSD mouse model study utilized a PTSD-like animal model to evaluate the aberrant astrocytic GABA homeostasis, confirming the consistency with humans. Genetic mouse model study investigated the potential benefits of genetically inhibiting MAOB to facilitate fear extinction in PTSD mouse models, suggesting implications for therapeutic strategies. PTSD mouse model study with KDS2010 treatment utilized a PTSD-like mouse model to evaluate the therapeutic efficacy of KDS2010, a reversible, selective MAOB inhibitor, in treating PTSD by modulating prefrontal GABA synthesis pathways. PTSD post-traumatic stress disorder, GABA gamma-aminobutyric acid, ¹H-MRS proton magnetic resonance spectroscopy, ASL arterial spin labeling, CBF cerebral blood flow, MAOB monoamine oxidase B, ABAT 4-aminobutyrate aminotransferase

Fig. 2

[Cross-sectional clinical study] PTSD-related prefrontal GABA alteration, CBF, and symptom severity. a Prefrontal and limbic ROIs for CBF measurement are depicted on an axial T1-weighted image. The right panels display color-coded mean CBF maps for the healthy control, persistent PTSD, and recovered PTSD groups. b Prefrontal GABA levels, assessed using ¹H-MRS, were elevated in the persistent PTSD group compared to healthy controls, while levels in the recovered PTSD group were similar to those of controls. c The persistent PTSD group exhibited reduced prefrontal CBF relative to healthy controls, whereas the recovered PTSD group showed prefrontal CBF comparable to the control group. d Both persistent and recovered PTSD groups showed higher limbic CBF levels compared to the healthy control group. e In the combined PTSD group (persistent and recovered), higher prefrontal GABA levels were associated with lower prefrontal CBF. f No significant association was found between prefrontal GABA and limbic CBF levels in the PTSD group. g Mediation analysis indicated that the link between elevated prefrontal GABA levels and greater PTSD symptom severity could potentially be mediated by reduced prefrontal CBF. h However, no mediation effect was observed for limbic CBF. Error bars in the graphs indicate standard errors of the mean. PTSD post-traumatic stress disorder, ROI region-of-interest, VOI voxel-of-interest, CBF normalized resting-state cerebral blood flow, GABA gamma-aminobutyric acid, ¹H-MRS proton magnetic resonance spectroscopy, MRI magnetic resonance imaging

Fig. 3

[Longitudinal clinical study] Prefrontal GABA normalization and PTSD symptom improvement in trauma-exposed individuals. a PTSD symptom severity, assessed by the CAPS scores, showed significant improvement from baseline to the 8-month follow-up. b Prefrontal GABA levels in the trauma-exposed group normalized to the levels observed in the trauma-unexposed group by the time of follow-up assessment. c Prefrontal CBF levels increased over time in the trauma-exposed group. d Limbic CBF levels in the trauma-exposed group showed a decreasing trend, but the change was not statistically significant. e In the trauma-exposed group, decreases in prefrontal GABA levels over time were correlated with an improvement in PTSD symptoms. Error bars in the graphs indicate standard errors of the mean. Abbreviations: CAPS Clinician-Administered Post-traumatic Stress Disorder Scale for DSM-5, FU follow-up, GABA gamma-aminobutyric acid, CBF normalized resting-state cerebral blood flow, PTSD post-traumatic stress disorder

Fig. 4

[Postmortem human brain study] Alterations in prefrontal astrocytic GABA in the postmortem brains from PTSD patients. a, b Representative images of S100β (a) and GABA (b) in the prefrontal cortices of postmortem brains from a control individual and a patient with PTSD. c Quantification of GABA intensity in the S100β-positive area between the control and PTSD groups. d, e Representative images of MAOB (d) and GFAP (e) in the prefrontal cortices of postmortem brains from a control individual and a patient with PTSD. f Quantification of MAOB intensity in the GFAP-positive area of the control and PTSD groups. g, h Representative images of ABAT (g) and GFAP (h) in the prefrontal cortices of postmortem brains from a control individual and a patient with PTSD. i Quantification of ABAT intensity in the GFAP-positive areas of the control and PTSD groups. j Quantification of the S100β-positive areas of the control and PTSD groups. k Quantification of the GFAP-positive areas of the control and PTSD groups. l Representative images from the Sholl analysis. m Comparison of the sum of the intersections between the control and PTSD groups. Detailed information about the statistical values is provided in Supplementary Table 8. Error bars in the graphs indicate standard errors of the mean. **P < 0.01; ***P < 0.001. PTSD post-traumatic stress disorder, S100β S100 calcium-binding protein B, GABA gamma-aminobutyric acid, MAOB monoamine oxidase B, GFAP glial fibrillary acidic protein, ABAT 4-aminobutyrate aminotransferase

Fig. 5

[PTSD mouse model study with KDS2010 treatment] Astrocytic changes and elevated GABA levels in the IL cortex of a PTSD-like mouse model. a The timeline for the PTSD-like mouse model and experiments. b Contextual fear conditioning, extinction, and extinction retrieval sessions in the control and PTSD-like mouse model. c, d Y-maze test results showing alternation percentage for spatial working memory (c) and total arm entries (d). e Representative confocal and Imaris images of S100β, MAOB, and GABA in the control and PTSD groups. A circle indicates the Sholl analysis. f, g Quantification of GABA (f) and MAOB (g) intensity in the S100β-positive area from confocal images. h Representative confocal and Imaris images of GFAP and ABAT in the control and PTSD groups. i Quantification of ABAT intensity in the GFAP-positive area. j Quantification of S100β-positive area. k-m The summary graph shows the ramification index (k), ending radius (I), and sum of intersections (m) in Sholl analysis. n Quantification of GFAP-positive area. o A schematic image of a whole-cell patch-clamp recording from an IL cortical neuron and representative traces of tonic GABA currents. p–r A summarized graph showing tonic GABA currents (p), sIPSC amplitude (q), and the frequency (r) in IL cortical neurons across the groups. s Schematic diagram of spike probability measurements and representative traces of evoked EPSPs from layer II stimulation (n = 3 mice per group). t Spike probability across stimulation intensities (50–800 μA) (left) and comparison of spike probability at 300 μA (right). Detailed information about the statistical values is provided in Supplementary Tables 9 and 10. Error bars in the graphs indicate standard errors of the mean. * P < 0.05; **P < 0.01; ***P < 0.001. ns not significant. PTSD post-traumatic stress disorder, SPS single prolonged stress, CBF cerebral blood flow, GABA gamma-aminobutyric acid; S100β, S100 calcium-binding protein B, MAOB monoamine oxidase B, ABAT 4-aminobutyrate aminotransferase, GFAP glial fibrillary acidic protein, sIPSC spontaneous inhibitory postsynaptic current, a.u. arbitrary unit, IL infralimbic, CC corpus callosum

Fig. 6

[Genetic mouse model study] Role of IL astrocytic MAOB in modulating extinction memory in a PTSD mouse model. a Astrocyte-specific gene-silencing of MAOB in the IL cortex, employing pSico-shRNA with GFAP-Cre virus. b Timeline for the development of shMAOB+PTSD and Sc+PTSD mouse models, followed by immunohistochemistry, electrophysiology, and contextual fear conditioning. c Contextual fear conditioning, extinction, and extinction retrieval sessions in the Sc+PTSD and shMAOB-PTSD groups. d, e Y-maze test results showing alternation percentage for spatial working memory (d) and total arm entries (e). f Immunostaining analyses and confocal microscopy images of GABA, MAOB, S100β, and NeuN in the Sc+PTSD and shMAOB+PTSD groups. g Quantification of the proportion of S100β-positive cells among mCherry-positive cells in Sc and shMAOB groups showing no significant difference. h, i Quantification of GABA (h) and MAOB (i) intensity within the mCherry-positive astrocytic areas across the groups. j Quantification of S100β-positive area. k Representative Sholl analysis images of astrocytes. l The summary graph shows the ramification index (left), ending radius (middle), and sum of intersections (right). m Representative traces of tonic GABA currents in the Sc + PTSD and shMAOB + PTSD groups. n-p Summarized data on tonic GABA currents (n), sIPSC frequency (o), and amplitude (p) of the IL cortical neuron across the groups. q Experimental timeline and schematic representation of MAOB-KO mice expressing GFAP::MAOB or GFAP::GFP (control) for astrocyte-specific expression. r Contextual fear conditioning results showing extinction and retrieval performance (left) and freezing behavior in the retrieval session (right) in MAOB-WT littermates, MAOB-KO + GFAP::GFP and MAOB-KO + GFAP::MAOB groups. s Representative images of S100β and GABA staining across groups, including Sholl analysis images for astrocytes. t, u Quantification of GABA (t) and S100β (u) intensity in S100β-positive areas. v Quantification of S100β-positive areas. w–y The summary graph shows the ramification index (w), ending radius (x), and sum of intersections (y) in Sholl analysis. Detailed information about the statistical values is provided in Supplementary Tables 11,12, and 13. Error bars in the graphs indicate standard errors of the mean. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant. AAV adeno-associated virus, Lenti lentivirus, IL infralimbic PL prelimbic, CG anterior cingulate CC corpus callosum, AP anterior-posterior, ML medial-lateral, DV dorsal-ventral, MAOB monoamine oxidase B, KD knockdown, KO knockout, WT wild type, PTSD post-traumatic stress disorder, SPS single prolonged stress, shRNA single hairpin RNA, GFAP glial fibrillary acidic protein, Sc scramble shRNA, shMAOB MAOB shRNA, GABA gamma-aminobutyric acid, mCh mCherry; a.u., arbitrary unit, S100β S100 calcium-binding protein B, sIPSC spontaneous inhibitory postsynaptic current

Fig. 7

[PTSD mouse model study with KDS2010 treatment] Effects of MAOB-dependent tonic GABA alterations on CBF and fear extinction in the PTSD mouse model and their reversal through pharmacological MAOB inhibition. a The timeline for the PTSD-like mouse model KDS2010 (10 mg/kg/day, ad libitum drinking) treatments, and behavior experiments. b Behavioral data from the contextual fear conditioning, extinction, and extinction retrieval in the control, PTSD-like mouse, and PTSD-like mouse+KDS2010 groups. c, d Y-maze test results showing alternation percentage for spatial working memory (c) and total arm entries (d). e Representative confocal and Imaris images of S100β, MAOB, and GABA in the control, PTSD, and PTSD + KDS2010 groups (arrows indicate the location of inset images). Circle indicates Sholl analysis of astrocytes. f, g Quantification of GABA (f) and MAOB (g) intensity in the S100β-positive areas. h Representative confocal images of GFAP and ABAT in the control, PTSD, and PTSD + KDS2010 groups. i Quantification of ABAT intensity in the GFAP-positive areas. j, k Quantification of S100β- (j) and GFAP- (k) positive area. l-n Quantification of ramification index (l), ending radius (m), and sum of intersection (n). o Representative traces of tonic GABA currents across the groups. p-r A summarized graph showing tonic GABA currents (p), sIPSC amplitude (q), and the frequency (r) in IL cortical neurons. s Schematic diagram of spike probability measurements and representative traces of evoked EPSPs (n = 3 mice per group). t Spike probability across stimulation intensities (50-800 μA) (left) and comparison of spike probability at 300 μA (right). u The timeline for the PTSD-like mouse model, LDF recording, and NMDA stimulation for CBF measurement (left). Representative image of LDF probe installation and NMDA application in the frontal cortex (middle, right). v A representative trace of NMDA stimulation-evoked CBF changes for each group. w–y Violin graphs depicting the peak amplitude (w), AUC (x), and time to peak (y) NMDA stimulation–evoked CBF changes in the control, PTSD, and PTSD + KDS2010 groups. Detailed information about the statistical values is provided in Supplementary Table 13. Error bars in the graphs indicate standard errors of the mean. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant. PTSD post-traumatic stress disorder, SPS single prolonged stress, CBF cerebral blood flow, NMDA, N-methyl-D-aspartate, LDF Laser Doppler Flowmetry, GABA gamma-aminobutyric acid, S100β S100 calcium-binding protein B, MAOB monoamine oxidase B, ABAT 4-aminobutyrate aminotransferase, GFAP glial fibrillary acidic protein, sIPSC spontaneous inhibitory postsynaptic current, a.u. arbitrary unit, AUC area under the curve, IL infralimbic, CC corpus callosum