Abstract
Astrocytes are closely linked to depression, and the prefrontal cortex
(PFC) is an important brain region involved in major depressive
disorder (MDD). However, the underlying mechanism by which astrocytes
within PFC contribute to MDD remains unclear. Using single-nucleus RNA
sequencing analyses, we show a significant reduction in astrocytes and
attenuated pleiotrophin-protein tyrosine phosphatase receptor type Z1
(PTN-PTPRZ1) signaling in astrocyte-to-excitatory neuron communication
in the PFC of male MDD patients. We find reduced astrocytes and PTN in
the dorsomedial PFC of male mice with depression induced by chronic
restraint and social defeat stress. Knockdown of astrocytic PTN induces
depression-related responses, which is reversed by exogenous PTN
supplementation or overexpression of astrocytic PTN. The antidepressant
effects exerted by astrocytic PTN require interaction with PTPRZ1 in
excitatory neurons, and PTN-PTPRZ1 activates the AKT signaling pathway
to regulate depression-related responses. Our findings indicate the
PTN-PTPRZ1-AKT pathway may be a potential therapeutic target for MDD.
Subject terms: Depression, Molecular neuroscience, Astrocyte
__________________________________________________________________
Astrocytes in the prefrontal cortex are closely linked to depression,
but the underlying mechanisms remain unclear. Here, the authors show
that reduction of astrocytic pleiotrophin in the dorsomedial prefrontal
cortex contributes to depression-like phenotype in male mice.
Introduction
Major depressive disorder (MDD) is a prevalent mental disorder
characterized by depressed mood, fatigue or loss of energy in daily
life, anhedonia, and suicidal ideation^[46]1. MDD is a leading cause of
disability worldwide, contributing significantly to the global burden
of disease^[47]2,[48]3. The increased incidence of suicide attempts and
suicide-related death contributes to increased mortality in populations
with MDD^[49]4.
The pathogenesis of MDD is thought to be a result of the combined
effects of genetic, environmental, psychological, and biological
factors^[50]1,[51]5. Multiple brain regions are involved in the
pathology of MDD, including the prefrontal cortex (PFC), anterior
cingulate cortex, hippocampus, amygdala, nucleus accumbens, and
anterior insula^[52]6–[53]8. The PFC is an important region related to
MDD^[54]9,[55]10. Positron emission tomography radioligand examination
has revealed an increased distribution of enzymes regulating
nonserotonergic monoamine metabolism within the PFC in patients with
major depressive episodes^[56]11. Previous studies have shown
structural and functional changes in the dlPFC after treatment in MDD
patients^[57]12–[58]14. In clinical practice, treatment measures
targeting the dlPFC lead to differences in the outcomes of the
antidepressant effect and somatic symptoms in MDD
patients^[59]15,[60]16. The medial prefrontal cortex (mPFC) in rodents
corresponds to the dorsolateral prefrontal cortex (dlPFC) in
primates^[61]17,[62]18. The role of the PFC in rodents in the
pathogenesis of depression has also been widely studied^[63]19–[64]22.
However, the molecular etiology of MDD is complex and remains unclear.
Astrocytes, which are ubiquitously distributed across the brain and
closely interact with neurons, play a crucial role in the maintenance
of normal functions in the central nervous system^[65]23. The important
role of astrocytes in the pathogenesis of MDD has been widely
studied^[66]24,[67]25. Astrocyte dysfunction in the mPFC contributes to
aberrant functional connectivity in depression-related networks in
mice^[68]26. Morphological changes in astrocytes are associated with
MDD progression and the severity of depression-related
behaviors^[69]27–[70]29. Astrocyte pathology, including aberrant
protein and mRNA expression of astrocyte markers such as glial
fibrillary acidic protein (GFAP) and S100β, has been reported to be
related to the development of MDD^[71]30. Over 25 years ago,
researchers identified a significant reduction in the density and size
of astrocytes within the dlPFC of depressed individuals^[72]31. Glial
ablation in the PFC via toxins specifically targeting astrocytes
efficiently induces depression-related behaviors in rats^[73]32. A
reduction in the density of GFAP+ astrocytes in the hippocampus has
also been observed in a mouse model of depression induced by chronic
mild stress^[74]33. Depletion of PFC GFAP+ cells has been reported to
induce anhedonia-like behavior but not anxiety-like deficits, which can
be reversed by activating GFAP+ cells in the PFC^[75]34. However, how a
reduction in the number of PFC astrocytes affects depression at the
molecular level has not yet been fully elucidated.
MDD is a neurological disease characterized by synaptic dysfunction
involving astrocytes, which are intimately involved in interactions
with neuronal synapses and play a role in the processing of synaptic
information^[76]35. Depression induced by chronic stress is closely
associated with excitation-inhibition imbalances within brain
regions^[77]36. Pleiotrophin (PTN), also known as heparin-binding
growth-associated molecule, is a developmentally regulated and secreted
growth factor. The PTN facilitates the maturation of newborn
neurons^[78]37, and depletion of the PTN may result in the loss of
neurons^[79]38. Abnormalities in neurogenesis and neuronal functions is
related to MDD^[80]1,[81]39, and astrocytes are involved in the
modulation of neuronal activity, synaptic function and neuronal
growth^[82]30. Previous studies have shown that transgenic mice with
PTN overexpression have more astrocytes in the hippocampus, which
suggests that the endogenous PTN level may be involved in the
modulation of astrocytic responses^[83]40,[84]41. In a study focused on
cell-cell communication in traumatic brain injury, astrocytes were
found to be the main source of upregulated PTN in the hippocampus
during the subacute phase, and elevated PTN promoted neurite
growth^[85]42. PTN derived from astrocyte have been reported to
contribute to improving recovery from acute neuroinflammation^[86]43.
However, whether the astrocytic PTN in the PFC participates in the
pathogenesis of MDD via the modulation of neural growth or synaptic
function remains unclear. A previous study revealed that a reduction in
PTN induces learning and memory impairment in mice, and the underlying
mechanism may involve interruption of the interaction between PTN and
protein tyrosine phosphatase receptor type Z1 (PTPRZ1), a receptor of
PTN, which further affects adult hippocampal neurogenesis by activating
AKT signaling^[87]44. In a study of the hippocampus of insomnia-induced
cognitively impaired mice, the expression of PTN was reduced, leading
to reduced binding to PTPRZ1 on the postsynaptic membrane^[88]45.
Abnormal regulation of PTN has been reported to be relevant to synaptic
dysfunction^[89]46,[90]47. The astrocytes and PTN in the dorsomedial
PFC (dmPFC) may affect the development of MDD via the modulation of
neurogenesis and synaptic function. However, whether astrocytic PTN
regulates MDD by modulating the density size and morphology of
astrocytes in the PFC remains unclear.
In this study, we observed that the astrocytic PTN in the dmPFC was
decreased in a stress-induced mouse model of depression. PTN activity
in astrocytes in the dmPFC was sufficient to alleviate
depression-related behaviors, increase neuronal excitability, and
enhance synaptic signaling transmission. The effects of the astrocytic
PTN in the dmPFC on depression required interaction with PTPRZ1 in
excitatory neurons, which further affected phosphorylated AKT
signaling. In conclusion, our findings reveal that the interaction of
the astrocytic PTN in the PFC with PTPRZ1 in excitatory neurons is a
potential target for MDD treatment.
Results
A single-nucleus transcriptomic atlas of the human prefrontal cortex in major
depressive disorder patients and psychiatrically healthy controls
Many studies have indicated a critical role of the prefrontal cortex in
the pathophysiology of major depressive disorder^[91]34. To elucidate
the cellular composition and potential mechanisms of MDD in this
region, we reanalyzed snRNA-seq data from Brodmann area 9 (BA9) in the
dlPFC of 17 male patients who died during an episode of MDD and the
other 17 from matched mentally healthy male individuals ([92]GSE144136,
Supplementary Fig. [93]1A, Supplementary Data [94]1). After quality
control filters (see “Methods”), a total of 71,565 nuclear
transcriptomes from the 34 brains were retained for subsequent
analysis, of which 39,801 nuclei originated from MDD patients and
31,764 nuclei from controls (Fig. [95]1A, B, Supplementary Fig. [96]1B,
and Supplementary Data [97]2). We conducted snRNA-seq data analysis
following the Seurat pipeline and integrated this dataset on the basis
of individual samples using the Harmony algorithm to correct for the
batch effect (see Methods). The uniform mixing of samples and groups
suggested effective correction of the batch effect. The nuclei were
classified into seven major cell types (Fig. [98]1A, C, Supplementary
Fig. [99]1C, and Supplementary Data [100]3) on the basis of canonical
marker expression, including excitatory neurons (n = 46,180),
identified by the expression of SATB2, SLC17A7, and CAMK2A; inhibitory
neurons (n = 13,641), marked by GAD1 and GAD2; astrocytes (Astro,
n = 3510), which were positive for GFAP; oligodendrocytes (Oligo,
n = 4662), marked by MBP; oligodendrocyte precursor cells (OPC,
n = 2139), which were positive for PCDH15; endothelial cells (Endo,
n = 330), which expressed CLDN5 and VTN; and microglia/macrophage
(Micro/Macro, n = 1103), defined by their classical markers CX3CR1 and
MRC1. Given the high-resolution advantage of single-nucleus data, we
further annotated the excitatory or inhibitory neuronal populations
into more refined subtypes on the basis of distinct gene expression
patterns, allowing for a more detailed characterization of cortical
cellular architecture (see “Methods” for a full list of markers for
neuron subclusters; Fig. [101]1D, E and Supplementary Data [102]3).
Overall, our clustering results were consistent with those of previous
studies of snRNA-seq analysis of the human prefrontal cortex^[103]48.
Fig. 1. Single-nucleus atlas of the human prefrontal cortex in major
depressive disorder patients and psychiatrically healthy controls.
[104]Fig. 1
[105]Open in a new tab
A UMAP plots of 71,565 single nuclei grouped into nonneurons, including
astrocytes, oligodendrocytes, oligodendrocyte precursor cells,
endothelial cells, microglia/macrophages, excitatory neurons and
inhibitory neurons. Each dot represents a single nucleus and is colored
according to the cell type. Astro astrocytes, Oligo oligodendrocytes,
OPC oligodendrocyte precursor cells, Endo endothelial cells,
Micro/Macro microglia/macrophage, Ex excitatory neurons, Inhib
inhibitory neurons, L layer (i.e., Ex_c1_L2_L4, excitatory neurons of
the c1 subtype from cortex layer 2 to layer 4). B UMAP plot of the
above single cells colored according to their group (left panel) or
sample origin (right panel). C Feature plots showing the normalized
expression of marker genes for each cell type. Each dot represents a
single cell, and the depth of color from gray to red represents low to
high expression. D, E Dot plot showing the normalized expression of
marker genes for each subtype of inhibitory (D) or excitatory neurons
(E). The depth of color from blue to yellow represents low to high
expression, and the size of the circles represents the proportion of
expression, increasing from small to large. F Group prevalence of cell
clusters estimated by the Ro/e score (34 snRNA-seq) (see “Methods”). G
Box plots showing the proportions of cell clusters in the suicide and
control groups (n = 34). The centerlines denote median values, and the
whiskers denote 1.5× the interquartile range. P values were calculated
using two-sided Wilcoxon tests and adjusted for multiple comparisons
via the Benjamini‒Hochberg correction method. H Bar plots showing the
proportion of each cell type in suicide or control patients. Each bar
represents an individual patient, with cell proportions coded as
different colors for cell types on the right. See Supplementary
Data [106]5 for statistical details, including the statistical test
used for data analysis and the exact P value. Source data are provided
as a Source Data file.
Astrocyte subsets may be key regulators of the pathophysiology of MDD
The degree of infiltration of each cell type may reflect their distinct
functional roles. To assess these differences, we utilized multiple
methods to compare the changes in cell proportions between the two
groups. Ro/e analysis revealed a significant decrease in the
proportions of astrocytes, endothelial cells, OPCs, and
oligodendrocytes in the MDD group, whereas the proportions of neurons
and Micro/Macros were modestly increased. Notably, the number of
astrocytes exhibited the greatest reduction (Fig. [107]1F). The results
of the Wilcoxon test with the Benjamini‒Hochberg procedure between
groups further confirmed the significant reduction in the proportion of
astrocytes in MDD patients (Astro, P[adj] = 0.0045; Endo,
P[adj] = 0.041; Excit_neuron, P[adj] = 0.093; Inhib_neuron,
P[adj] = 0.81; Micro/Macro, P[adj] = 0.81; Olig, P[adj] = 0.20; OPC,
P[adj] = 0.093; Fig. [108]1G, Supplementary Data [109]4 and [110]5). In
addition, the cell composition of each sample demonstrated a consistent
decrease in astrocytes across all MDD samples compared with controls,
eliminating the influence of outliers (Fig. [111]1H).
MDD is a disease closely related to genetic background^[112]1. To
investigate how these genetic variants in relevant cell types mediate
the disease at the single-cell level, we performed an integrated
analysis of GWAS data and single-cell data using scPagwas algorithm
(see “Methods”, Supplementary Data [113]5). We found that astrocytes
with higher trait-relevant scores (TRSs) showed the strongest
enrichment in MDD patients (Supplementary Fig. [114]2A). Consistently,
astrocytes were significantly associated with MDD at the cell type
level (astro P[adj] = 1.11 × 10^−11; Supplementary Fig. [115]2B, and
Supplementary Data [116]5).
Taken together, astrocytes not only showed the most significant change
in cell proportions but were also identified as the key cell type by
which genetic variants influence MDD, supporting their critical role in
the pathophysiology of MDD.
Astrocytes markedly influence excitatory neurons in MDD, and their
interaction via PTN pathway signaling is significantly reduced
To explore intercellular communication and discover major signaling
changes in response to disease, we performed CellChat analysis
(Supplementary Fig. [117]2A). We found that in both the MDD and control
groups, astrocytes, as sender cells, interact with mainly excitatory
neurons, inhibitory neurons, OPCs, and oligodendrocytes in the human
cortex (Fig. [118]2A). The largest communication change was observed
between astrocytes and excitatory neurons (Fig. [119]2B). Furthermore,
by comparing the overall probability of communication between
astrocytes and excitatory neurons in the cortex of the MDD and control
groups, we found that five pathways were observed to be active in MDD,
including the NCAM, NEGR, EPHA, NGL, and SEMA3 (Fig. [120]2C and
Supplementary Fig. [121]3C) pathways. Among these pathways, 16 out of
29 signaling pathways showed decreased activity, including seven
classic pathways related to neurotrophic support and neuromodulation,
such as the PTN, PSAP, and SEMA series (Fig. [122]2C, Supplementary
Fig. [123]3C, E–H).
Fig. 2. Cell‒cell interaction network generated by CellChat showing that the
PTN pathway is significantly decreased in the suicide group compared with the
control group.
[124]Fig. 2
[125]Open in a new tab
A Circle plot showing the number of interactions between astrocytes
(source) and various target cell populations in the two groups. B
Heatmap showing the differential number of interactions or interaction
strengths among different cell populations across the two groups. C Bar
chart showing the comparison of the overall information flow of each
signaling pathway or ligand‒receptor pair between astrocytes and
excitatory neurons between groups. Significant signaling pathways were
ranked on the basis of differences in the overall information flow
within the inferred networks between the two groups. Paired Wilcoxon
tests were performed to assess the statistical significance. Pathways
enriched in the control group are highlighted in red on the y-axis,
whereas those enriched in the suicide group are in blue. D Circle plot
showing the PTN signaling pathway from astrocytes (outgoing arrows) to
all excitatory neuron subtypes (ingoing arrows) across the suicide and
control groups. The circle sizes are proportional to the number of
cells in each cell group, and the edge weights represent interaction
strength. A thicker edge indicates a stronger signaling interaction. E
Dot plot showing decreased ligand‒receptor interactions between
astrocytes and excitatory neuron subtypes in control (red) or suicide
(blue) samples. Ligand‒receptor interactions are indicated in columns.
The means of the average expression levels of two interacting molecules
are indicated by the color heatmap (bottom panel), with blue to red
representing low to high expression. P values were obtained via
one-sided permutation tests and are indicated by the circle size. F
Violin plot showing the gene expression distributions of PTN signaling
pathway-related key ligands in astrocytes and receptors in
excit_neurons between the control (red) and suicide (blue) groups of
astrocytes and excitatory neuron subtypes. P values were derived from
two-sided Wilcoxon tests and adjusted for multiple comparisons via the
Benjamini‒Hochberg correction method. See Supplementary Data [126]5 for
statistical details, including the statistical test used for data
analysis and the exact P value. Source data are provided as a Source
Data file.
Astrocytic PTN in the hippocampus has been reported to be related to
the regulation of neurogenesis in traumatic brain injury^[127]42, and
aberrant PTN expression may impair synaptic function^[128]46. In the
depression model, we investigated the interaction between astrocytic
PTN and molecules in excitatory neurons. Specific to the PTN signaling
pathway, CellChat revealed that its related ligand‒receptor pairs,
PTN‒PTPRZ1, PTN‒SDC3, PTN‒NCL, and PTN‒ALK, were also significantly
decreased in the communication from Astro to Ex_L2‒L6 (Fig. [129]2E and
Supplementary Fig. [130]3D), corresponding to the PTN overall pathway
change (Fig. [131]2D). Finally, the gene expression of key ligands in
astrocytes and receptors in excitatory neurons related to the PTN
signaling pathway was also significantly lower in the MDD group than in
the control group, which is consistent with previous findings on
pathways and ligand‒receptor pairs (Astro PTN, P[adj] = 0.0043;
Excit‒neuron PTPRZ1, P[adj] = 5.63 × 10^−23; SDC3,
P[adj] = 1.18 × 10^−38; SDC4, P[adj] = 4.16 × 10^−18; NCL,
P[adj] = 2.23 × 10^−32; ALK, P[adj] = 0.78; Fig. [132]2F, Supplementary
Fig. [133]1D, [134]4A, B, and Supplementary Data [135]5). In summary,
CellChat analysis suggested that the decreased activity of the PTN
signaling pathway in astrocyte-to-excitatory neuron communication may
contribute to MDD pathogenesis.
The density of astrocytes and PTN expression in the dmPFC are decreased in a
mouse model of depression
The above results suggest that the proportion of astrocytes and the
expression of the astrocytic PTN in the human dlPFC are associated with
the development of MDD. In terms of developmental origin, the human
dlPFC corresponds to the dmPFC in rodents. To investigate the
underlying mechanism of MDD, we prepared a wild-type (WT) C57BL/6 mouse
model of depression under chronic restraint stress (CRS) (Fig. [136]3A)
and chronic social defeat stress (Supplementary Fig. [137]7A), which
successfully induced depression-related behaviors, including
depressive-like behaviors in the FST, TST, SPT, and SIT, and
anxiety-like behaviors in the OFT and EPM (Fig. [138]3B and
Supplementary Fig. [139]7B). S100β, SOX9, and GFAP were chosen as
markers for astrocytes. The WB and IHC results revealed that S100β,
SOX9, and GFAP expression were lower in the mice subjected to CRS
(Fig. [140]3C–H) or CSDS (Supplementary Fig. [141]7C–H) than in the
control mice. These results suggest that the reduction in the number of
astrocytes in the dmPFC is related to depression. To verify the
CellChat analysis that astrocytic PTN may contribute to depression,
additional in vivo experiments were performed. The mRNA (Fig. [142]3I)
and protein expression (Fig. [143]3J) of PTN in the dmPFC were
decreased in the CRS mice, in accordance with the IHC images
(Fig. [144]3K). The costaining of S100β with PTN, SOX9, and PTN, GFAP,
and PTN demonstrated that PTN was expressed in astrocytes in the dmPFC
(Fig. [145]3L). To specifically validate the expression level of PTN in
astrocytes, we isolated astrocytes from the dmPFC via an Anti-ASCA-2
MicroBeads Kit and subjected them to WB. Astrocytic PTN was
significantly reduced in the CRS mice (Fig. [146]3M), similar to that
in the CSDS mice (Supplementary Fig. [147]7I). Next, we explored
neuronal excitability in the dmPFC. Both the amplitude and frequency of
mEPSCs were significantly reduced in the depressed mice
(Fig. [148]3N–P), indicating diminished presynaptic and postsynaptic
transmission across synapses, which suggests attenuated synaptic
plasticity. Moreover, the firing frequency was decreased in mice with
depression (Fig. [149]3Q), demonstrating reduced neuronal excitability.
The expression of postsynaptic density protein 95 (PSD95) was also
reduced in CRS mice (Fig. [150]3R). The results of the sparse labeling
test revealed that the density of dendritic spines was significantly
lower in CRS mice than in control mice (Fig. [151]3S). These results
demonstrate that the reduction of astrocytes and PTN expression in the
dmPFC possibly contributes to the development of depression by reducing
neuronal excitability and inducing synapse impairments.
Fig. 3. CRS leads to a decreased proportion of astrocytes and PTN expression
in the dmPFC in mice.
[152]Fig. 3
[153]Open in a new tab
A Schematic of the experimental design (WT mice). B Analyses of
behavioral tests, including the OFT, EPM, FST, TST, and SPT (n = 10
mice per group). C–E Quantification of S100β (C), SOX9 (D), and GFAP
(E) expression by WB analyses (n = 6 mice per group). F–H IHC showing
reduced S100β + (F), SOX9+ (G), and GFAP+ (H) cells in the CRS mice
(n = 6 replicates). I, J RT‒PCR and WB analyses showing that PTN mRNA
(I) and PTN protein expression (J) were reduced in CRS mice (n = 6 mice
per group). K IHC showing reduced PTN+ cells in the dmPFC of CRS mice
(n = 6 replicates). L Representative images of PTN+ (red) and
S100β + (green) co-stained cells (upper panel), PTN+ (green) and SOX9+
(red) co-stained cells (middle panel), PTN+ (green) and GFAP+ (red)
co-stained cells (lower panel) (n = 6 replicates). M WB analysis
showing reduced PTN within isolated astrocytes in the CRS mice (n = 6
mice per group). N–P Traces of mEPSCs in CaMKIIα+ neurons from the
indicated groups (N), and quantification of the amplitude (O) and
frequency (P) between groups (n = 30 cells from 10 mice per group). Q
Traces of action potentials in CaMKIIα+ neurons and the corresponding
statistical analysis (n = 30 cells from 10 mice per group). R
Quantification of PSD95 expression (n = 6 mice per group). S Traces of
dendritic spines in CaMKIIα+ neurons and quantification of the density
(n = 30 cells from 10 mice per group). Diagrams of mouse CRS model and
mouse brain in (A) were created in BioRender. chi, d. (2025)
[154]https://BioRender.com/x64m191. All data are presented as
mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001. See Supplementary
Data [155]5 for statistical details, including the statistical test
used for data analysis and the exact P value. Source data are provided
as a Source Data file.
Knockdown of PTN in the dmPFC promotes depression-related behaviors, reduces
neuronal excitability, and leads to synapse impairments
Next, we bilaterally injected CMV-PTN-shRNA or CMV-PTN-shRNA NC into
the dmPFC of WT mice to further investigate the effect of PTN on
depression (Fig. [156]4A). The following experiments were performed
three weeks after the application of the AAVs. The knockdown efficiency
was verified through PCR, WB, and IHC analyses (Fig. [157]4B–D), which
revealed that both the mRNA and protein expression of PTN in the dmPFC
were significantly downregulated. In the behavioral tests, the mice
exhibited much poorer performance in both depressive-like behaviors
(FST, TST, SPT) and anxiety-like behaviors (OFT, EPM) after PTN was
knocked down in the dmPFC (Fig. [158]4E), indicating that PTN in the
dmPFC plays a crucial role in the modulation of depression-related
behaviors. Using CMV-PTN-shRNA and CaMKIIα-EGFP labeling, the
excitability of excitatory neurons in the dmPFC was tested in WT
C57BL/6 mice. Both the amplitude and frequency of mEPSCs were reduced
in the dmPFC PTN-knockdown mice (Fig. [159]4F–H), indicating that
presynaptic and postsynaptic excitatory transmission were decreased
when the dmPFC PTN was downregulated. In addition, knocking down dmPFC
PTN expression significantly reduced the firing frequency
(Fig. [160]4I). Given that the development of MDD is closely associated
with neuronal synapses and synaptic information processing, we
investigated how altered PTN expression affects neuronal synaptic
morphology. As shown in Fig. [161]4J, a significant reduction in the
density of dendritic spines was observed after the inhibition of PTN
expression in the dmPFC. WB analysis revealed that PSD95 expression was
attenuated after PTN was knocked down in the dmPFC of the mice
(Fig. [162]4K). The results revealed that the knockdown of the PTN in
the dmPFC led to synaptic impairments. Taken together, these findings
suggest that PTN deficiency in the dmPFC facilitates depression,
possibly through attenuated neuronal excitability and synaptic
dysfunction.
Fig. 4. Knockdown of the PTN in the dmPFC promotes depression-related
behaviors, reduces neuronal excitability, and leads to synaptic impairment.
[163]Fig. 4
[164]Open in a new tab
A Schematic of the experimental design (WT mice). B–D RT‒PCR (B, n = 6
mice per group), WB (C, n = 6 mice per group), and IHC (D, n = 6
replicates) analyses showing the knockdown efficiency of AAV-PTN-shRNA.
E Analyses of behavioral tests, including the OFT, EPM, FST, TST, and
SPT (n = 10 mice per group). F–H Traces of mEPSCs in CaMKIIα+ neurons
from the indicated groups (F), and the statistical analysis of the
amplitude (G) and frequency (H) between groups (n = 30 cells from 10
mice per group). I Traces of action potentials in CaMKIIα+ neurons from
the indicated groups and the corresponding statistical analysis (n = 30
cells from 10 mice per group). J Traces of dendritic spines from
CaMKIIα+ neurons and quantification of the density of dendritic spines
(n = 30 cells from 10 mice per group). K Quantification of PSD95
expression (n = 6 mice per group). Diagrams of the mouse injection
model and mouse brain in (A) were created in BioRender. chi, d. (2025)
[165]https://BioRender.com/x64m191. All data are presented as
mean ± SEM. **P < 0.01, ***P < 0.001. See Supplementary Data [166]5 for
statistical details, including the statistical test used for data
analysis and the exact P value. Source data are provided as a Source
Data file.
Astrocytic-specific knockdown of the PTN in the dmPFC facilitates
depressive-like responses
The results of the snRNA-seq analysis (Figs. [167]1 and [168]2)
revealed that among the seven major cell types, astrocytes had the
strongest association with MDD, and the PTN signaling pathway in
astrocyte-to-excitatory neuron communication was linked to MDD
pathogenesis. Co-stained imaging revealed that PTN is expressed in
astrocytes in the dmPFC (Fig. [169]3L). To determine whether PTN
derived from astrocytes affects MDD modulation, we utilized
intraperitoneal injection of tamoxifen (TAM) and bilateral injection of
DIO-PTN-shRNA into the dmPFC of Aldh1l1-Cre/ERT2 mice to specifically
knock down astrocytic PTN (Fig. [170]5A). We separately examined the
effects of tamoxifen on WT mouse behavior, and the results showed no
significant changes (Supplementary Fig. [171]9C). To verify the
knockdown efficiency of AAV transfection, astrocytes were isolated from
dmPFC tissue of Aldh1l1-Cre/ERT2 mice and examined using PCR and WB,
which revealed that both the mRNA and protein expression of PTN were
reduced (Fig. [172]5B, C). In the behavioral studies, the astrocytic
PTN-knockout Aldh1l1-Cre/ERT2 mice presented depression-related
behaviors, including depressive-like behaviors and anxiety-like
behaviors (Fig. [173]5D). After specifically knocking down the
astrocytic PTN, both the amplitude and frequency of mEPSCs
(Fig. [174]5E–G), as well as the firing frequency (Fig. [175]5H), were
reduced in PTN-shRNA-treated mice. Moreover, downregulation of the
astrocytic PTN also resulted in diminished dendritic spine density in
dmPFC neurons (Fig. [176]5I) and decreased PSD95 expression
(Fig. [177]5J). These findings suggest that the astrocyte-derived PTN
in the dmPFC contributes to depression modulation by affecting neuronal
excitability and synaptic functions.
Fig. 5. Astrocytic PTN in the dmPFC modulates depression-related behaviors
and neuronal excitability and synaptic functions.
[178]Fig. 5
[179]Open in a new tab
A Schematic of the experimental design (Aldh1l1-Cre/ERT2 mice). B, C
Validation of the knockdown efficiency of DIO-PTN-shRNA in
Aldh1l1-Cre/ERT2 mice through RT‒PCR analysis of PTN mRNA (B) and WB
analysis of PTN protein (C) in astrocytes isolated from the dmPFC
(n = 6 mice per group). D Analyses of behavioral tests, including the
OFT, EPM, FST, TST, and SPT (n = 10 mice per group). E–G Traces of
mEPSCs in CaMKIIα+ neurons in Aldh1l1-Cre/ERT2 mice from the indicated
groups (E), and the statistical analysis of the amplitude (F) and
frequency (G) of mEPSCs between groups (n = 30 cells from 10 mice per
group). H Traces of action potentials in CaMKIIα+ neurons in
Aldh1l1-Cre/ERT2 mice from the indicated groups and the corresponding
statistical analysis (n = 30 cells from 10 mice per group). I Traces of
dendritic spines from CaMKIIα+ neurons in Aldh1l1-Cre/ERT2 mice and
quantification of the density (n = 30 cells from 10 mice per group). J
Quantification of PSD95 expression (n = 6 mice per group). Diagrams of
the mouse injection model and mouse brain in (A) were created in
BioRender. chi, d. (2025) [180]https://BioRender.com/x64m191. All data
are presented as mean ± SEM. **P < 0.01, ***P < 0.001. See
Supplementary Data [181]5 for statistical details, including the
statistical test used for data analysis and the exact P value. Source
data are provided as a Source Data file.
Exogenous PTN supplementation or astrocytic PTN upregulation in the dmPFC
exerts antidepressant effects
Since the depletion of astrocytic PTN in the dmPFC leads to
depression-related behaviors, we investigated whether the exogenous
administration of PTN into the dmPFC could elicit antidepressant-like
effects. A cannula, positioned with its tip in the dmPFC of WT C57BL/6
mice for drug administration, was implanted prior to preparing the
depression mouse model. The mice then received either PTN or control
PBS injected into the dmPFC for further observation (Fig. [182]6A). The
results showed that the infusion of exogenous PTN into the dmPFC
effectively improved depression-related behaviors in stress-induced
depressed mice (Fig. [183]6B). Similar experiments were conducted in
depressed WT mice induced by CSDS, and similar results revealed that
exogenous PTN supplementation also relieved the depression-related
behaviors induced by CSDS (Supplementary Fig. [184]8A). With respect to
neuronal excitability, exogenous PTN infusion significantly increased
the amplitude and frequency of mEPSCs (Fig. [185]6C–E) and increased
the firing frequency (Fig. [186]6F). In addition, increased dendritic
spine density (Fig. [187]6G) and PSD95 expression levels (Fig. [188]6H)
were detected following PTN infusion in the dmPFC of WT CRS mice.
Fig. 6. PTN infusion alleviates depression-related behaviors in
CRS-susceptible mice.
[189]Fig. 6
[190]Open in a new tab
A Schematic of the experimental design (WT mice). B Analyses of
behavioral tests, including the OFT, EPM, FST, TST, and SPT (n = 10
mice per group). C–E Traces of mEPSCs recorded in CaMKIIα+ neurons (C),
and the statistical analysis of the amplitude (D) and frequency (E)
between groups (n = 30 cells from 10 mice per group). F Traces of
action potentials in CaMKIIα+ neurons from the indicated groups and the
corresponding statistical analysis (n = 30 cells from 10 mice per
group). G Traces of dendritic spines from CaMKIIα+ neurons and
quantification of the density (n = 30 cells from 10 mice per group). H
Quantification of PSD95 expression (n = 6 mice per group). I Schematic
of the experimental design (Aldh1l1-Cre/ERT2 mice). J Validation of AAV
transfection efficiency via quantification of PTN expression in
isolated astrocytes (n = 6 mice per group). K Analyses of behavioral
tests, including the OFT, EPM, FST, TST, and SPT (n = 10 mice per
group). L–N Traces of mEPSCs in CaMKIIα+ neurons from the indicated
groups (L), and quantification of the amplitude (M) and frequency (N)
(n = 30 cells from 10 mice per group). O Traces of action potentials in
CaMKIIα+ neurons from the indicated groups and the corresponding
statistical analysis (n = 30 cells from 10 mice per group). P Traces of
dendritic spines in CaMKIIα+ neurons from the indicated groups and
quantification of the density (n = 30 cells from 10 mice per group). Q
Quantification of PSD95 expression (n = 6 mice per group). Diagrams of
mouse injection model, CRS model, and mouse brain in (A, I) were
created in BioRender. chi, d. (2025)
[191]https://BioRender.com/x64m191. All data are presented as
mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001. See Supplementary
Data [192]5 for statistical details, including the statistical test
used for data analysis and the exact P value. Source data are provided
as a Source Data file.
In order to clarify the effects of endogenous PTN upregulation in the
dmPFC on depression-related responses, we constructed AAV-DIO-PTN OE to
overexpress astrocytic PTN in Aldh1l1-Cre/ERT2 mice (Fig. [193]6I). The
efficiency of the constructed AAVs was verified by isolation of
astrocytes and quantification of the astrocytic PTN through WB
(Fig. [194]6J). As expected, overexpression of astrocytic PTN in the
dmPFC of the Aldh1l1-Cre/ERT2 mice also sufficiently improved
depression-related behaviors in both the CRS model (Fig. [195]6K) and
the CSDS model (Supplementary Fig. [196]8B). Moreover, astrocytic PTN
overexpression in the dmPFC of the Aldh1l1-Cre/ERT2 mice significantly
increased the amplitude and frequency of mEPSCs (Fig. [197]6L–N), as
well as the firing frequency in the CRS model mice (Fig. [198]6O).
Greater dendritic spine density (Fig. [199]6P) and increased PSD95
expression (Fig. [200]6Q) were also observed following astrocytic PTN
overexpression in the dmPFC in the CRS model mice. These findings
reveal that increased astrocytic PTN expression in the dmPFC may
alleviate depression-related responses.
Interaction between astrocytic PTN and excitatory-neuronal PTPRZ1 in the
dmPFC mediates antidepressant effects
Next, we explored how the expression of astrocytic PTN affects other
cells and affects depression-related responses. PTN exerts various
effects by binding to several putative transmembrane receptors, such as
PTPRZ1, ALK, and syndecan 3 (SDC3)^[201]44. Previous CellChat analysis
(Fig. [202]2E) indicated that PTPRZ1, SDC3, SDC4, NCL, or ALK might
interact with the PTN between astrocytes and excitatory neurons,
contributing to MDD pathogenesis. Thus, we constructed corresponding
AAVs to knock down PTPRZ1, SDC3, SDC4, NCL, or ALK in CaMKIIα+
excitatory neurons in the dmPFC of WT mice to evaluate their role in
depression. The results revealed that the downregulation of SDC3, SDC4,
NCL, or ALK in excitatory neurons did not elicit depression-related
behaviors (Supplementary Fig. [203]6A–D). We then focused on the
potential effects of PTPRZ1 on depression. According to previous
studies, PTPRZ1 is most likely the membrane receptor through which PTN
influences neural function^[204]44,[205]45. Therefore, we investigated
whether astrocytic PTN modulates antidepressant effects through its
interaction with PTPRZ1 in excitatory neurons. We first examined the
level of PTPRZ1 in WT mice and found that both the mRNA and protein
expression of PTPRZ1 within the dmPFC were lower in CRS mice than in
control mice (Fig. [206]7A–C). PTPRZ1 expression in isolated neurons
was also significantly reduced in CRS mice (Fig. [207]7D) and CSDS mice
(Supplementary Fig. [208]7J). We applied CaMKIIα-Cre and
DIO-PTPRZ1-shRNA to the dmPFC to knock down the expression of PTPRZ1 in
excitatory neurons in WT mice (Fig. [209]7E). To confirm the knockdown
efficiency of AAV transfection, we utilized a neuron isolation
MicroBeads kit to extract neurons from dmPFC tissue and confirmed that
both the mRNA and protein expression of PTPRZ1 were significantly
reduced (Fig. [210]7F, G). Behavioral tests revealed that
downregulation of PTPRZ1 in excitatory neurons also induced
depression-related behaviors, as the experimental mice exhibited worse
performance in the OFT, EPM, FST, TST, and SPT (Fig. [211]7H).
Knockdown of PTPRZ1 in excitatory neurons led to decreased neuronal
excitability and synaptic impairment, as evidenced by the reduced
amplitude and frequency of mEPSCs (Fig. [212]7I–K), lower firing
frequency (Fig. [213]7L), decreased dendritic spine density
(Fig. [214]7M), and diminished PSD95 expression (Fig. [215]7N). These
findings suggest that PTPRZ1 in excitatory neurons may also exert
antidepressant effects.
Fig. 7. Astrocytic PTN within dmPFC interacts with PTPRZ1 in excitatory
neurons to mediate antidepressant effects.
[216]Fig. 7
[217]Open in a new tab
A–C Quantification of PTPRZ1 mRNA (A) and protein (B) (n = 6 WT mice
per group), and verification by IHC analysis (C) (n = 6 replicates). D
Quantification of isolated neuronal PTPRZ1 (n = 6 WT mice per group). E
Schematic of the experimental diagram (WT mice) and IHC showing PTPRZ1
(green) co-expressed with CaMKIIα + (red) neuron (n = 6 replicates). F,
G Validation of DIO-PTPRZ1-shRNA through quantification of PTPRZ1 mRNA
(F) and protein (G) (n = 6 mice per group). H Behavioral tests
analyses, including OFT, EPM, FST, TST, and SPT (n = 10 mice per
group). I–K Traces of mEPSCs in CaMKIIα+ neurons (I), and
quantification of amplitude (J) and frequency (K) (n = 30 cells from 10
mice per group). L Traces and quantification of action potentials in
CaMKIIα+ neurons (n = 30 cells from 10 mice per group). M Traces and
quantification of dendritic spines from CaMKIIα+ neurons (n = 30 cells
from 10 mice per group). N Quantification of PSD95 (n = 6 mice per
group). O Schematic of experimental design (WT mice). P Behavioral
tests analyses, including OFT, EPM, FST, TST, and SPT (n = 10 mice per
group). Q–S Traces of mEPSCs in CaMKIIα+ neurons (Q), and
quantification of amplitude (R), and frequency (S) (n = 30 cells from
10 mice per group). T Traces and quantification of action potentials
from CaMKIIα+ neurons (n = 30 cells from 10 mice per group). U Traces
and quantification of dendritic spines from CaMKIIα+ neurons (n = 30
cells from 10 mice per group). V Quantification of PSD95 (n = 6 mice
per group). Diagrams of the mouse brain and injection model in (E, O)
were created in BioRender. chi, d. (2025)
[218]https://BioRender.com/x64m191. All data are presented as
mean ± SEM. N.S. = not significant, *P < 0.05, **P < 0.01,
***P < 0.001. See Supplementary Data [219]5 for statistical details,
including the statistical test used for data analysis and the exact P
value. Source data are provided as a Source Data file.
Next, we investigated the interaction between the astrocytic PTN and
PTPRZ1 in excitatory neurons in the dmPFC. After specifically knocking
down PTPRZ1 in excitatory neurons, exogenous PTN or vehicle was
administered to the dmPFC of WT mice (Fig. [220]7O). We found that the
depletion of PTPRZ1 induced depression-related behaviors in WT mice,
whereas subsequent exogenous PTN administration caused no significant
changes in these behaviors (Fig. [221]7P). Moreover, the administration
of PTN failed to restore neuronal excitability (Fig. [222]7Q–T) or
enhance neuronal synaptic transmission (Fig. [223]7U, V) when PTPRZ1 in
excitatory neurons was depleted. In addition, we selectively
overexpressed astrocytic PTN in the dmPFC of Aldh1l1-Cre/ERT2 mice
after PTN was knocked down with CMV-PTN-shRNA, and the Aldh1l1-Cre/ERT2
mice greatly recovered from depression-related behaviors with the
subsequent upregulation of astrocytic PTN (Supplementary Fig. [224]9A,
B). The Co-IP experiment was conducted to demonstrate the interaction
between PTN and PTPRZ1 (Supplementary Fig. [225]9D). Overall, the
results revealed that the astrocytic PTN in the dmPFC interacts with
PTPRZ1 in excitatory neurons to mediate antidepressant effects.
Astrocytic PTN in the dmPFC interacts with PTPRZ1 in excitatory neurons to
mediate antidepressive effects by activating the AKT signaling pathway
To investigate the downstream molecular mechanisms underlying
PTN-PTPRZ1 signaling in MDD, we next conducted a NicheNet analysis, a
cell‒cell communication tool that can predict the responsive target
genes involved in altered ligand‒receptor interactions (Supplementary
Fig. [226]5A–E). The NicheNet results predicted 92 target genes
following the PTN-PTPRZ1 pair (Supplementary Fig. [227]5E). Pathway
enrichment of these target genes revealed that the downstream pathways
were related to kinase activity, kinase binding, and activation of
protein kinase activity (Fig. [228]8A). Furthermore, GSEA of DEGs in
excitatory neurons, the effector cells executing pathway functions,
revealed a decrease in PI3K-AKT-MTOR signaling in the suicide group
(Fig. [229]8B). Previous studies have reported that PTN regulates adult
neurogenesis by binding to PTPRZ1 to activate the AKT signaling
pathway^[230]44. Collectively, these results indicate that the AKT
signaling pathway may serve as a downstream target of PTN-PTPRZ1
activity in astrocyte-to-excitatory neuron communication.
Fig. 8. Astrocytic PTN interacts with PTPRZ1 in excitatory neurons to mediate
antidepressant effects by activating the AKT signaling pathway.
[231]Fig. 8
[232]Open in a new tab
A Bar plot of the GO pathway enrichment analysis showing the
downregulated signaling pathways of target genes predicted from
NicheNet. B GSEA results of excitatory DEGs showing the enrichment of
PI3K_AKT_MTOR_signaling was attenuated in the suicide group. C
Quantification of p-AKT and total AKT expression (n = 6 WT mice per
group). D Schematic of experimental design (WT mice). E Analyses of
behavioral tests, including the OFT, EPM, FST, TST, and SPT (n = 10
mice per group). F–H Traces of mEPSCs recorded from CaMKIIα+ neurons
(F), and quantification of the amplitude (G) and frequency (H) (n = 30
cells from 10 mice per group). I Traces and quantification of action
potentials from CaMKIIα+ neurons (n = 30 cells from 10 mice per group).
J Traces of dendritic spines from CaMKIIα+ neurons and quantification
of the density (n = 30 cells from 10 mice per group). K Schematic of
the experimental design (WT mice). L, M Quantification of p-AKT and
total AKT expression (n = 6 mice per group). N Schematic of
experimental design (WT mice). O Analyses of behavioral tests,
including the OFT, EPM, FST, TST, and SPT (n = 10 mice per group). P–R
Traces of mEPSCs recorded from CaMKIIα+ neurons (P), and quantification
of the amplitude (Q), and frequency (R) (n = 30 cells from 10 mice per
group). S Traces and quantification of action potentials from CaMKIIα+
neurons (n = 30 cells from 10 mice per group). T Traces of dendritic
spines from CaMKIIα+ neurons and quantification of the density (n = 30
cells from 10 mice per group). Diagrams of the mouse, mouse brain and
CRS model in (D, K, N) were created in BioRender. chi, d. (2025)
[233]https://BioRender.com/x64m191. All data are presented as
mean ± SEM. N.S. = not significant, **P < 0.01, ***P < 0.001. See
Supplementary Data [234]5 for statistical details, such as the
statistical test used for data analysis and the exact P value. Source
data are provided as a Source Data file.
We first examined the level of AKT within the dmPFC in WT mice
subjected to CRS. WB analysis revealed that the level of phosphorylated
AKT (p-AKT) was significantly reduced in the CRS mice, whereas no
obvious differences in total AKT were detected (Fig. [235]8C). We
applied SC79, an activator of AKT, to the dmPFC of WT CRS mice
(Fig. [236]8D). SC79 significantly improved depression-related
behaviors (Fig. [237]8E), increased the amplitude and frequency of
mEPSCs (Fig. [238]8F–H), potentiated the firing rates of action
potentials (Fig. [239]8I), and increased the density of dendritic
spines (Fig. [240]8J), demonstrating the potential antidepressant
effects of the AKT pathway. Similar results in depression-related
behavioral tests were found in the WT CSDS mouse model (Supplementary
Fig. [241]8C). Next, we explored the relationship between the
PTN-PTPRZ1 pathway and AKT signaling. To examine whether PTPRZ1
depletion or exogenous PTN infusion affects AKT signaling, we conducted
a series of experiments with WT mice (Fig. [242]8K). WB analysis
revealed that exogenous PTN infusion in the dmPFC increased p-AKT,
whereas PTPRZ1 knockdown in excitatory neurons reduced p-AKT expression
(Fig. [243]8L, M). However, administering PTN in the dmPFC failed to
alter p-AKT expression when PTPRZ1 in excitatory neurons had already
been inhibited (Fig. [244]8L, M). We then applied MK2206, an inhibitor
of AKT, to the dmPFC of WT mice (Fig. [245]8N). MK2206 administration
induced depression-related behaviors (Fig. [246]8O), reduced the
amplitude and frequency of mEPSCs (Fig. [247]8P–R), decreased the
firing rates of action potentials (Fig. [248]8S), and decreased the
density of dendritic spines (Fig. [249]8T). These results suggest that
PTN-PTPRZ1 in the dmPFC modulates depression responses by activating
the AKT signaling pathway.
Discussion
In this study, through snRNA-seq data and CellChat analyses, we found
that the proportion of astrocytes in the dlPFC was decreased and that
the PTN signaling pathway involved in astrocyte-to-excitatory neuron
communication was attenuated in male MDD patients compared with
mentally healthy controls. In vivo experiments revealed that reduced
numbers of astrocytes and diminished PTN expression in the dmPFC were
detected in a male mouse model of depression induced by CRS and CSDS.
Furthermore, PTN from astrocytes in the dmPFC modulated
depression-related responses. The knockdown of astrocytic PTN induced
depression-related behaviors, reduced neuronal excitability, and
impaired synaptic function, which was reversed by supplementation with
exogenous PTN or the upregulation of astrocytic PTN in the dmPFC.
Moreover, astrocytic PTN exerted antidepressant effects via interaction
with PTPRZ1 in excitatory neurons and further activation of the AKT
signaling pathway.
Astrocytes play a crucial role in the pathology of MDD, as astrocytes
regulate glutamatergic neurotransmission, modulate neural and synaptic
plasticity through the secretion of growth factors such as
brain-derived neurotrophic factor (BDNF), and mediate
neuroinflammation^[250]49. Postmortem studies indicate that both the
frequency and severity of reactive astrogliosis are reduced in the
brains of individuals with MDD^[251]50. Our results also revealed that
the density of astrocytes in the dmPFC was significantly reduced in a
mouse model of depression induced by CRS and CSDS. We utilized S100β,
SOX9, and GFAP as the markers for astrocytes, and all markers
demonstrated a reduction of astrocyte number in the dmPFC of mice under
chronic stress. Andrea et al. reported that unpredictable chronic mild
stress followed by social isolation increased GFAP immunoreactivity and
increased s100β-positive cell density in the dentate gyrus of the
hippocampus in mice^[252]51. This discrepancy is likely due to the
different mouse models and different brain regions used in the two
studies. In addition, in Andrea’s study^[253]51, the tendency of
GFAP-positive cell immunoreactivity and cell density in the PFC was not
in accordance with that in the dentate gyrus. Astrocytes play a vital
role in regulating the transmission of glutamate, and glutamate serves
as a key messenger facilitating communication between central neurons
and astrocytes^[254]52. Repair of impaired astrocyte function can
affect antidepressant effects by ensuring neuronal activity^[255]53.
Mice that have diminished expression of the glutamate transporter GLT-1
in astrocytes of the habenula are more prone to stress and show
heightened immobility in tail-suspension tests^[256]54. These
observations support the importance of preserving the function of
astrocytes in MDD treatment.
PTN, also known as heparin-binding growth-associated molecule, is an
extracellular matrix protein that plays important roles in neurite
outgrowth, synaptogenesis, and synaptic plasticity^[257]55,[258]56. PTN
is closely associated with neuroplasticity, and PTN can mediate newborn
neuron development through activation of the AKT signaling
pathway^[259]37. PTN can exert protective effects on the development of
addictive and neurodegenerative diseases such as Parkinson’s disease
and Alzheimer’s disease^[260]57. The pathogenesis of MDD is related to
impaired excitation‒inhibition balance within neurons in the central
nervous system^[261]58. The clearance of glutamate, which affects
neuronal excitability and the strength of glutamatergic synapses,
contributes to depression^[262]59. As the main excitatory
neurotransmitter released by synapses, glutamate is involved in
synaptic plasticity, cognitive processes, and reward and emotional
processes. Stress can affect the presynaptic secretion of glutamate and
activate downstream signaling pathways by binding to receptors on the
postsynaptic membrane^[263]60. The impairment of glutamatergic neuron
function is involved in the development of depression, and the
regulation of the glutamatergic system affects depression-related
responses^[264]61,[265]62. Astrocytes are capable of adjusting the
strength of synapses by enhancing spontaneous excitatory postsynaptic
currents^[266]63. This finding is consistent with our findings in the
present study. The knockdown of the astrocytic PTN in the dmPFC results
in a decreased firing frequency in excitatory neurons, a reduced
amplitude and frequency of mEPSCs, diminished dendritic spine density,
and decreased PSD95 expression, indicating attenuated neuronal
excitability and hypoactive synaptic functions. The reduced astrocytic
PTN in the dmPFC induced depression, possibly due to the impairment of
neuronal excitability and glutamatergic transmission.
As a member of the transmembrane protein tyrosine phosphatase family,
PTPRZ1 belongs to the PTN receptor family and binds to PTN through its
extracellular domain^[267]64. The role of PTN-PTPRZ1 in several central
nervous system diseases has been studied^[268]65,[269]66. Potentiation
of the PTN-PTPRZ1 pathway in the hippocampus activates AKT signaling,
resulting in the restoration of neurogenesis and alleviation of memory
deficits in aging mice^[270]44. PTN-PTPRZ1 signaling promotes
oligodendrocyte precursor cell proliferation and differentiation during
developmental myelination and remyelination after injury. PTPRZ1 is
also highly expressed in glial cells of the central nervous system,
including astrocytes and oligodendrocytes^[271]65. However, whether
PTPRZ1 in astrocytes affects the development of MDD in the same way as
it does in excitatory neurons in our study is not yet clear. The
astrocytic PTN in the dmPFC may mediate the progression of MDD via
interactions with not only PTPRZ1 in excitatory neurons but also PTPRZ1
in astrocytes; however, more experiments are needed to clarify these
effects. The role of the AKT pathway in depression has been widely
studied. The regulation of downstream AKT signaling facilitates the
amelioration of depression-related behaviors^[272]67,[273]68. The
neurotrophic factor brain-derived neurotrophic factor (BDNF) can
mediate the AKT pathway to improve synaptic plasticity and exert
antidepressant effects^[274]69. These observations are consistent with
our finding that the activation of phosphorylated AKT by the PTN-PTPRZ1
interaction can exert antidepressant effects.
There are some limitations in our study. First, the current study
focused on male MDD patients and male mice, mainly because the
prevalence, symptoms, and molecular signatures of MDD differ by
sex^[275]70–[276]72 and we think it is more reasonable to analyze these
factors separately. The effect of PTN-PTPRZ1 pathway on female subjects
with depression needs further verification. Second, direct evidence
demonstrating the interaction between astrocytic PTN and neuronal
PTPRZ1 was not presented in this study, and more experiments are needed
to confirm that interaction in the future. Third, we found that the
phosphorylation of AKT was affected by the aberrant expression of
astrocytic PTN and neuronal PTPRZ1 and that interfering with AKT
activity through exogenous administration of an AKT activator or
inhibitor mediated the depression-related responses. However, the
intervention measures we utilized lacked cell type specificity. Fourth,
Li X et al. reported that tamoxifen caused negative effects on the
behaviors of mice^[277]73, whereas our results showed that the
administration of tamoxifen had no significant effects on the observed
behaviors. The discrepancy may be due to differences in the total
dosage of tamoxifen administration between the two studies. As the main
plasma active metabolite of tamoxifen, 4-hydroxytamoxifen can
effectively shorten the time to exert effects and reduce the frequency
of drug administration, which can possibly reduce the side effects of
tamoxifen to a greater extent. Using 4-hydroxytamoxifen might be a
better choice for future experiments.
In summary, astrocytic PTN in the dmPFC functions as a critical
regulator of depression-related behaviors in a male mouse model of
depression induced by CRS and CSDS. Astrocytic PTN in the dmPFC
interacts with PTPRZ1 in the CaMKIIα+ excitatory neurons to modulate
neuronal excitability and synaptic transmission, which further
activates the AKT signaling pathway to exert antidepressant effects.
These findings highlight the PTN-PTPRZ1-AKT axis in
astrocyte-to-excitatory neuron communication as a promising therapeutic
target for MDD.
Methods
Dataset acquisition and study participants
The processed MDD single-nucleus RNA sequencing data from GBM samples
([278]GSE144136) were obtained from the Gene Expression Omnibus (GEO,
[279]http://www.ncbi.nlm.nih.gov/geo/) database, and the GWAS summary
statistics for MDD patients (170,756 cases and 329,443 controls) were
obtained from the IEU OpenGWAS ([280]https://gwas.mrcieu.ac.uk/)
database with a GWAS ID of ieu-b-102. The published MDD cohort^[281]48
consisted of 17 male patients and controls. Cases met the diagnostic
criteria for MDD and died by suicide, whereas individuals in the
control group died by accident (n = 6) or natural causes (n = 11).
Postmortem brain samples were dissected from Brodmann area 9 (dlPFC),
and snRNA sequencing was performed (Supplementary Data [282]1).
Single-cell data processing and cell clustering
We downloaded the processed snRNA sequencing data from the GEO dataset,
which had been aligned and quantified using the Cell Ranger Single-Cell
Software Suite (10X Genomics Cellranger v2.1.0). The gene expression
data were also mapped to the human genome (GRCh38-1.2.0). For the
quality check procedure, we first removed the cluster cells originally
labeled “Mix” from the dataset. Next, any cells were removed if they
expressed fewer than 201 genes or if more than 20% of the UMIs were
linked to mitochondrial genes^[283]74,[284]75. To eliminate batch
effects, the R package Harmony (version 1.2.0) was used to remove batch
correction before clustering analysis. Other steps, including Seurat
object creation, data normalization, data scaling, identification of
variable genes, dimensional reduction, clustering, and UMAP projection,
were performed following the standard pipeline of the Seurat R package
(version 4.4.0, [285]https://satijalab.org/seurat).
Cluster annotation
All the nuclei were annotated as distinct major cell types on the basis
of the average gene expression of known canonical marker genes,
including excitatory neurons (markers: SATB2, SLC17A7, and CAMK2A),
inhibitory neurons (GAD1 and GAD2), astrocytes (GFAP, AQP4, GLUL, SOX9,
GJA1, NDRG2, and ALDH1A1), oligodendrocytes (MBP, PLP1, MOBP, and MAG),
oligodendrocyte precursor cells (PCDH15, PDGFRA, OLIG1, OLIG2, and
PTGDS), endothelial cells (CLDN5 and VTN), and microglia/macrophage
(CX3CR1, MRC1, SPI1, and TMEM119) cells. In addition, SNAP25, STMN2,
and RBFOX3 were used as common neuron markers across all the neuronal
populations.
We further identified subclusters and annotated them as specific cell
subtypes within excitatory and inhibitory neurons on the basis of the
average expression of the corresponding gene sets. The marker genes for
each subcluster within the major cell types were identified using the
FindAllMarkers function in the Seurat package. For excitatory neurons,
subclusters were annotated according to layer-specific gene markers as
follows: layers II–IV (CUX2, RASGRF2, and PVRL3), layer IV/V (RORB),
layer V (SULF2, PCP4, and HTR2C), layers V/VI (TOX, ETV1, RXFP1, and
FOXP2), and layer VI (NR4A2, SYNPR, TLE4, and NTNG2). For inhibitory
neurons, subclusters were annotated on the basis of well-known markers,
including CCK, CALB, VIP, SST, PVALB, SLC32A1, and DLX.
Differential expression analysis
Analysis of differentially expressed genes (DEGs) for key ligands and
receptors between the suicide and control groups was performed using
two-sided Wilcoxon tests within each responding cluster, with
Benjamini–Hochberg correction applied to adjust P values for multiple
testing. A BH-adjusted P value less than 0.05 was considered
statistically significant. Additionally, the jjVolcano function from
the scRNAtoolVis package (version 0.0.7) was used to visualize DEGs in
a Manhattan plot between the suicide and control groups across all
clusters.
Group preference of each cell type
To quantify the cell type enrichment across groups, we compared the
observed to expected cell numbers in each type according to the
following formula as we described previously^[286]76,[287]77:
[MATH:
Ro/e=observedexpected :MATH]
1
where the expected cell numbers for each combination of cell types and
groups are obtained via the chi-square test. Ro/e is the ratio of
observed to expected cell numbers in each cell type. Ro/e > 1 suggests
a preference for this cell type in this group, and Ro/e < 1 suggests
that cells of the given cell type are observed with less frequency than
random expectations in the specific group.
Cell‒cell interaction analysis
CellChat
To compare differences in cell‒cell communication networks across
different sources, we utilized CellChat (version 1.6.1,
[288]https://github.com/sqjin/CellChat)^[289]78. The normalized
expression data, preprocessed using Seurat, were input into CellChat
for downstream analysis. We applied standard preprocessing functions
with default parameters, including identifyOverExpressedGenes,
identifyOverExpressedInteractions, and projectData, to identify
significantly overexpressed genes and interactions and project the data
onto inferred interaction networks. We subsequently ran the core
functions computeCommunProb, computeCommunProbPathway, and aggregateNet
using default settings to compute the communication probabilities and
aggregate the interaction networks. The MDD and control groups were
merged for further analysis using computeNetSimilarityPairwise,
allowing pairwise comparisons of network differences. To visualize the
communication patterns, the functions compareInteractions, rankNet, and
netVisual_bubble were used.
Initially, we performed CellChat analysis of all the nuclei and
determined that excitatory neurons were most regulated by astrocytes.
We then subset the relevant nuclei from excitatory neurons and
astrocytes for a more focused CellChat analysis followed by the
abovementioned steps.
NicheNet
We used NicheNet (version 2.1.7,
[290]https://github.com/saeyslab/nichenetr)^[291]79 to infer
interactions between astrocytes and excitatory neurons, designating
astrocytes as “Sender” cells and excitatory neurons as “Receiver”
cells. Only genes expressed in more than 10% of the cells within
clusters were considered for further ligand‒receptor interaction
analysis. We extracted ligands with upregulated activity that
overlapped with the CellChat results from the control vs. suicide
comparison (i.e., ligands with downregulated activity in the suicide
group compared with the control group) and the top 1000 targets from
the differentially expressed genes of the “Sender” and “Receiver” cells
for paired ligand‒receptor activity analysis. The
ligand_activity_target_heatmap function was applied to visualize ligand
regulatory activity, with activity scores ranging from 0 to 1. In
addition, a heatmap was generated to illustrate the expression of
differentially expressed ligands and receptors, which was calculated by
averaging gene expression in the relevant cell populations and scaling
across the indicated clusters.
Pathway enrichment analysis
Gene ontology (GO) analysis
To perform GO analysis, the potential target genes of excitatory
neurons regulated by astrocytes through the PTN pathway, as predicted
by NicheNet, were analyzed using the online tool Metascape^[292]80
([293]http://metascape.org/gp/index.html) to identify the enriched
pathways. Only gene sets in the “GO Biological Processes”, “GO
Molecular Functions” and “GO Cellular Components” categories were
considered, and a P value less than 0.05 was considered statistically
significant.
Gene set enrichment analysis (GSEA)
To explore the potential downstream signaling pathways affected by PTN
signaling, we conducted GSEA using GSEABase (version 1.64.0) and gene
sets from the Molecular Signatures Database
([294]https://www.gsea-msigdb.org/gsea/msigdb/human/collections.jsp)^[2
95]81. The DEGs in excitatory neurons between the suicide and control
groups were subjected to GSEA. Genes were ranked by log2-fold change
from the differential expression analysis, GSEA was conducted using
hallmark (H) and curated (C2) gene sets with 1000 permutations, and a P
adjusted value less than 0.05 was considered statistically significant.
Pathway-based polygenic regression (scPagwas) to identify MDD-relevant
clusters
To identify the genetic associations of cell clusters with MDD, we used
scPagwas ([296]https://dengchunyu.github.io/)^[297]82, which uses an
optimized polygenic regression model to integrate snRNA-seq and GWAS
data. For GWAS data preparation, we downloaded the GWAS summary
statistics for MDD from the IEU OpenGWAS database and conducted data
pruning following the scPagwas guidelines, including GWAS data
reprocessing, SNP extraction, plink, filtering of LD information and
integration of the result files for chromosomes 1–22 with default
parameters. For snRNA-seq preparation, we subset the normalized and
scaled expression data from the suicide group. The prepared GWAS and
snRNA-seq data were then input into scPagwas for downstream analysis,
with 1000 top genes and 200 iterations specified.
Animals
Male C57/BL6 mice (8–10 weeks) were purchased from the Institute of
Experimental Animals of Guangdong Medicine Experimental Animal Center.
Male Aldh1l1-Cre/ERT2 (Cat# 031008) mice (8–10 weeks) were purchased
from the Jackson Laboratory. All the mice were housed on a standard
12-h light/dark cycle (light from 7 a.m. to 7 p.m.) at a constant
temperature of 25 ± 1 °C and 50% humidity, with food and water
available at all times. Ten mice per group were used in behavioral
tests, electrophysiological recording, and dendritic spine analysis.
Six mice per group were used in RT-PCR, western blotting (WB), and
immunohistochemistry (IHC). All experimental procedures were approved
by the Use Committee of Sun Yat-sen University Cancer Center and the
Animal Care Committee (No. L102012024120D, No. L102012024228P) and were
conducted in accordance with the guidelines of the National Institutes
of Health (NIH).
Chronic restraint stress (CRS) model
The CRS model was established on the basis of procedures reported
previously^[298]83 with slight modifications. The mice were placed in
well-ventilated 50-mL plastic tubes without food or water from 9:00 AM
to 15:00 PM for 14 consecutive days. The plastic tubes prevented the
mice from moving freely or turning around but did not squeeze the
animals or cause any pain or physical injury. After being restrained,
the mice were sent back to their home cages immediately. The mice that
were not subjected to restraint were kept in their usual home cages
without any restrictions for 14 days.
Chronic social defeat stress (CSDS) model and social interaction test (SIT)
CSDS was performed as followed^[299]84. Adult male CD-1 mice (4–6
months old) were used as aggressors. Before each defeat, aggressors
were screened for aggressive behavior for 3 consecutive days. Two days
before the start of defeat, the CD-1 mice were housed on one side of a
perforated Plexiglass partition. During 10 consecutive days of CSDS,
experimental mice (7–8 weeks old) were subjected to direct physical
interaction with a CD-1 mouse for 10 min per day, and for the rest of
the day were placed on the other side of the Plexiglass divider,
allowing for sensory but not direct physical contact. The experimental
mice were exposed to a new CD-1 aggressor every day for 10 days.
Before the SIT, the mice were placed in the behavioral suite and
allowed to adapt for 1 h. After each test, the behavioral suite was
cleaned with 75% alcohol to prevent odor interference. SIT was
performed 24 h after CSDS modeling in a new CD-1 mouse. The test lasted
for 5 min. The experimental mice were allowed to explore freely in the
field (44 × 44 cm) in the first 2.5 min, where a rectangular container
(10 × 6 cm) was placed. The CD-1 mice were placed in the container
2.5 min later and observed in the “interaction area” (14 × 26 cm). The
statistical formula for the social interaction ratio was as follows:
time in the interaction area when there was a CD-1 target mouse/time in
the interaction area when there was no CD-1 target mouse × 100%.
Virus microinjection
Under continuous isoflurane inhalation anesthesia, the mice were placed
in a stereotaxic frame (RWD Life Science Co., Ltd.), and the body
temperature was maintained at 36 °C with a heating pad. The mice were
then treated with a bilateral stereotaxic injection of the virus into
the dmPFC (anteroposterior (AP), +1.7 mm; mediolateral (ML), ±0.4 mm;
and dorsoventral (DV), −2.3 mm relative to bregma). The virus was
infused bilaterally through a microinjector with a 33 G needle. The
virus mixture (150 nl) was infused for 10 min. After infusion, the
needle was retained at the injection site for an additional 10 minutes
before being withdrawn slowly. The mice were allowed to recover for 3
weeks to allow stable transgene expression.
The shRNA target sequences used in the study are listed in
Supplementary Table [300]1. CMV-PTN-shRNA was constructed to knock down
the expression of PTN, and CMV-PTPRZ1-shRNA was constructed to knock
down the expression of PTPRZ1. AAV-DIO-PTN-shRNA was used to
specifically knock down the expression of astrocytic PTN in the dmPFC
in Aldh1l1-Cre/ERT2 mice, with the assistance of intraperitoneally
administered tamoxifen (Sigma‒Aldrich, dissolved in corn oil,
75 mg/kg/day for 5 consecutive days) to ensure the activation of Cre
recombinase. CaMKIIα-Cre along with AAV-DIO-PTPRZ1-shRNA was applied to
specifically knock down the expression of PTPRZ1 in dmPFC excitatory
neurons. We constructed a DIO-PTN-Overexpression (DIO-PTN OE) to
overexpress the astrocytic PTN in the dmPFC in Aldh1l1-Cre/ERT2 mice.
All Aldh1l1-Cre/ERT2 mice received a 10-day rest for recovery after the
consecutive administration of tamoxifen.
For sparse labeling of CaMKIIα+ excitatory neurons in the dmPFC, the
mice were bilaterally injected with a total of 400 nL of viral cocktail
(1:1) of rAAV-CaMKIIα-FLP-WPRE-pA and rAAV-nEf1α-FDIO-EYFP-WPRE-pA.
Exogenous chemicals administration
For infusions into the dmPFC, mice received 1 µl of 5 µg/ml PTN
(MedChemExpress). For microinjections of AKT activator SC79 or AKT
inhibitor MK2206, 1 µl of 1 mg/ml SC79 (MedChemExpress; dissolved in
dimethyl sulfoxide in saline) or 1 µl of 1 mg/ml MK2206
(MedChemExpress; dissolved in dimethyl sulfoxide in saline) was
administered.
Behavioral tests
Open field test (OFT)
The mice were placed in an open chamber (50 × 50 × 40 cm), which was
made of gray polyvinyl chloride. The mice were allowed to freely
explore the apparatus. The mice were gently placed in the center, and
movement was recorded for 5 min. The total distance traveled and time
spent in the center (25 × 25 cm) were recorded by a video tracking
system and analyzed with software (Shanghai Jiliang Software
Technology, Co., Ltd.)
Elevated plus maze (EPM)
The apparatus consisted of two opposing open arms (35 × 5 cm) and two
opposing enclosed arms (35 × 5 × 15 cm), which were connected by a
central platform (5 × 5 cm) and positioned 50 cm above the ground. The
activity of the experimental subject within the apparatus was tracked
for 5 min via an overhead digital video camera. The time spent in the
open arms was recorded and analyzed.
Forced swimming test (FST)
The FST was performed in a glass cylinder (height 30 cm, diameter
12 cm), which was filled to 25 cm with water (22–24 °C). The mice were
placed in the cylinder for 6 min. The first 2 min were for the mice to
acclimate, and the duration of immobility was recorded during the last
4 min. The immobility time was recorded and analyzed with software
(Shanghai Jiliang Software Technology, Co., Ltd.).
Tail-suspension test (TST)
The mice were suspended by the tail 50 cm above the floor. The activity
was automatically monitored during the last 4 minutes of the 6-min test
with a threshold defining immobility behavior. The latency to the first
immobilization was also recorded.
Sucrose preference test (SPT)
The mice were acclimated to two bottles (50-ml tubes with fitted
ball-point sipper tubes) for 24 h. One bottle was filled with 1%
sucrose, and the other was filled with drinking water. Every 12 h, the
bottle positions were swapped to prevent position bias. After another
24 h, the amounts of sucrose and water consumed were recorded, and
sucrose preference (%) was calculated as sucrose consumption/(sucrose +
water consumption) × 100%.
RT-PCR
TRIzol was used to extract total RNA, and reverse transcription was
performed following the protocol of the polymerase chain reaction (PCR)
production kit (Accurate Biology, AG 11706). The primer sequences of
the investigated mRNAs for the PCR assay are shown in Supplementary
Table [301]2. The reaction cycle conditions were as follows: initial
denaturation at 95 °C for 3 min, followed by 40 cycles of 10 s at
95 °C, 20 s at 58 °C, and 10 s at 72 °C. The ratio of mRNA expression
in the dmPFC tissues was analyzed via the 2^−ΔΔCT method.
Western blotting (WB) analysis
Mouse brain tissues were lysed in ice-cold lysis buffer (RIPA buffer
with proteinase inhibitor) for 30 min and centrifuged at 10,000 × g for
20 min at 4 °C. Protein samples were separated by SDS‒PAGE and
transferred onto polyvinylidene difluoride (PVDF) membranes
(Millipore). The membranes were blocked with 5% milk in TBST for 1 h at
room temperature and then incubated with primary antibodies against
S100β (1:1000, Proteintech, Cat# 66616), SOX9 (1:1000, Proteintech,
Cat# 55152), GFAP (1:1000, Proteintech, Cat# 60190), PTN (1:1000,
Proteintech, Cat# 27117), PTPRZ1 (1:1000, Proteintech, Cat# 55125),
PSD95 (1:1000, Affinity, Cat# AF5283-50), AKT (1:1000, CST, Cat# 4685),
p-AKT (Ser473) (1:1000, CST, Cat# 4060) and GAPDH (1:1000, Fdbio
Science, Cat# FD0063) overnight at 4 °C. Information on the antibodies
used in the study can be found in Supplementary Table [302]3. On the
next day, the membranes were washed with TBST three times and incubated
with horseradish peroxidase (HRP)-conjugated secondary antibodies
(Fdbio Science) for 1 h at room temperature. Protein abundance was
quantified by analyzing the western blot bands using ImageJ software.
Quantified band intensities were normalized to GAPDH levels.
Unprocessed scans of the blots are provided in the Source Data file.
Immunohistochemistry (IHC)
The mice were anesthetized with phenobarbital and perfused with 4% PFA
in 0.1 M PBS. The brain tissues were postfixed overnight in 4% PFA at
4 °C and transferred to 30% sucrose in 0.1 M PBS. The brain tissues
were cut into 20 μm-thick sections on a freezing microtome. Then, the
sections were washed with PBS three times and incubated first in a
blocking buffer containing 3% bovine serum albumin in 0.2% Triton
X-100/PBS for 2 h at room temperature and then with primary antibodies
against PTN (1:200, Proteintech, Cat# 27117), PTPRZ1 (1:100,
Proteintech, Cat# 55125), S100β (1:200, Proteintech, Cat# 66616), SOX9
(1:200, Invitrogen, Cat# 14-9765-80), and GFAP (1:200, Proteintech,
Cat# 60190), in blocking buffer overnight at 4 °C. After that, the
sections were incubated with secondary antibody (Cy3 and Alexa fluor
488) at 37 °C for 60 min. The stained sections were examined using with
a Nikon confocal microscope equipped, and images were captured with a
Nikon DS-Qi2 camera.
Isolation of astrocytes and neurons from the mouse brain
Astrocytes and neurons were isolated from the dmPFC of mice using
magnetic-activated cell sorting (MACS)^[303]23. In brief, after the
mice were anesthetized with pentobarbital sodium and perfused with
ice-cold sterile PBS, the dmPFC tissue was dissociated at 37 °C for
15 min using the Adult Brain Dissociation Kit (Miltenyi Biotec, Cat#
130-107-677) with a gentleMACS Dissociator (Miltenyi Biotec, Cat#
130-093-235).
To isolate astrocytes, the cells were incubated with FcR Blocking
Reagent after the cell pellet was passed through 70-μm nylon mesh.
Next, the cells were incubated with Anti-ACSA-2 MicroBeads (Miltenyi
Biotec, Cat# 130-097-679) and then subjected to MACS through the LS
column.
To isolate neurons, after the brain tissue was dissociated, the cell
pellet was treated with cold debris removal solution and red blood cell
removal solution. Following the instructions of the Adult Neuron
Isolation Kit (Miltenyi Biotec, Cat# 130-126-603), the cells were
incubated with the Adult Non-Neuronal Cell Biotin-Antibody Cocktail and
Anti-Biotin MicroBeads and then subjected to MACS through the LS
column. The specificity of the Anti-ACSA-2 MicroBeads kit and Adult
Neuron Isolation Kit was validated and is shown in Supplementary
Fig. [304]9E, F.
Electrophysiological recording
The mice were anesthetized with isoflurane, and the whole brain was
quickly dissociated into prechilled and oxygenated dissection fluid
containing (in mM) 213 sucrose, 10 glucose, 26 NaHCO[3], 3 KCl, 1
NaH[2]PO[4]·2H[2]O, 10 MgCl[2], and 0.5 CaCl[2]. Acute brain slices
(300 µm) containing the dmPFC were acquired in chilled dissection fluid
using a microtome (VT1000S, Leica). The sections were transferred to
the incubation chamber and immersed in artificial cerebrospinal fluid
(in mM; 125 NaCl, 26 NaHCO[3], 5 KCl, 1.2 NaH[2]PO[4], 2.6 CaCl[2], 1.3
MgCl[2], and 10 glucose) at 30 °C for 1 h. After incubation, the
sections were placed in the slice chamber for electrophysiological
recording with continuous perfusion of artificial cerebrospinal fluid
(saturated with 95% O[2]/5% CO[2]).
The dmPFC excitatory neurons were identified by CaMKIIα-EGFP and
visualized for recording using an infrared-differential interference
contrast microscope (Olympus, Japan). We used patch pipettes (4–8 MΩ,
WPI, USA) for the whole-cell patch clamp recordings. The signal was
amplified by a MultiClamp 700B amplifier (Molecular Devices, USA). The
miniature excitability postsynaptic current (mEPSC) was sampled in the
presence of tetrodotoxin (1 μM) and picrotoxin (100 μM) at –70 mV. The
patch pipettes were filled with an intracellular solution containing
(in mM) 122 potassium-gluconate, 5 NaCl, 2 MgCl[2], 0.3 CaCl[2], 10
HEPES, 5 EGTA, 4 Mg-ATP, and 0.3 Na[3]-GTP. The action potentials were
recorded using the injected current (from 0 pA to 120 pA in 20 pA
increments, 500 ms duration) without any synaptic transmission
blockers. We calculated the number of action potentials induced by each
injected current (shown as firing rate–injected current (f–I) curves),
frequency (f–I), and amplitude of the mEPSCs using pCLAMP10.7.
Dendritic spine analysis
The mice were bilaterally injected with a total of 400 nL of a viral
cocktail (1:1) of rAAV-CaMKIIα-FLP-WPRE-PA or
rAAV-nEf1α-FDIO-EYFP-WPRE-PA to label CaMKIIα+ excitatory neurons in
the dmPFC. We used laser confocal photography to photograph dendritic
spines, acquired multilayer Z axes, and reconstructed the dendritic
spine image. The spines were classified into one of four morphological
subtypes: filopodial, thin, stub, or mushroom-shaped. ImageJ was used
([305]http://rsbweb.nih.gov/ij) to calculate the density of thin, stub,
and mushroom-shaped dendritic spines. Approximately ten randomly
selected neurons were analyzed per condition across two coverslips. The
density of spines was scored in dendritic segments 10 μm in length.
Finally, we counted and used the number of dendritic spines per 10 μm
to describe the density of the dendritic spines.
Statistical analysis
The statistical methods used for bioinformatics analysis in this paper
are indicated in the legend of each figure. For experimental
validation, SPSS 21.0 (IBM) was used to analyze the data, which are
presented as the means ± SEMs. The normality of the data was examined
via the Shapiro‒Wilk test. The normally distributed data were tested by
a two-sided unpaired t test, one- or two-way analysis of variance
(ANOVA). The nonnormally distributed data were analyzed by the Wilcoxon
test. Statistical significance was set at *P < 0.05, **P < 0.01,
***P < 0.001, N.S. not significant.
Reporting summary
Further information on research design is available in the [306]Nature
Portfolio Reporting Summary linked to this article.
Supplementary information
[307]Supplementary information^ (2MB, pdf)
[308]41467_2025_57924_MOESM2_ESM.docx^ (12.7KB, docx)
Description of Additional Supplementary Files
[309]Supplementary Data 1^ (11.4KB, xlsx)
[310]Supplementary Data 2^ (13KB, xlsx)
[311]Supplementary Data 3^ (2.9MB, xlsx)
[312]Supplementary Data 4^ (4.6MB, xlsx)
[313]Supplementary Data 5^ (34.5KB, xlsx)
[314]Reporting Summary^ (1.6MB, pdf)
[315]Transparent Peer Review file^ (1.2MB, pdf)
Source data
[316]Source Data^ (20.4MB, xlsx)
Acknowledgements