Skip to main content
NCBI home page
iScience logoLink to iScience
. 2019 Jun 11;17:24–35. doi: 10.1016/j.isci.2019.06.014

Chemogenetic Activation of Prefrontal Cortex in Shank3-Deficient Mice Ameliorates Social Deficits, NMDAR Hypofunction, and Sgk2 Downregulation

Luye Qin 1, Kaijie Ma 1, Zhen Yan 1,2,∗
PMCID: PMC6599088  PMID: 31247448

Summary

Haploinsufficiency of the SHANK3 gene is causally linked to autism spectrum disorders (ASDs) in human genetic studies. Here we found that chemogenetic activation of pyramidal neurons in the prefrontal cortex (PFC) of Shank3-deficient mice with the hM3D (Gq) DREADD restored social preference behaviors and elevated glutamatergic synaptic function in PFC. Moreover, the expression of Sgk2 (serum- and glucocorticoid-inducible kinase 2), a member of the Sgk family, which plays a key role in regulating the membrane trafficking of glutamate receptors, was diminished by Shank3 deficiency and rescued by Gq DREADD activation of PFC. Blocking Sgk function in Shank3-deficient mice prevented Gq DREADD from rescuing social and synaptic deficits, whereas blocking Sgk function in wild-type mice led to the attenuation of PFC glutamatergic signaling and the induction of autism-like social deficits. These results have provided a potential circuit intervention and molecular target for autism treatment.

Subject Areas: Neuroscience, Behavioral Neuroscience, Molecular Neuroscience, Cellular Neuroscience

Graphical Abstract

graphic file with name fx1.jpg

Open in a new tab

Highlights

  • •

    Chemogenetic activation of PFC rescues social deficits in Shank3-deficient mice

  • •

    Gq DREADD restores NMDAR and elevates AMPAR signaling in PFC of Shank3-deficient mice

  • •

    Gq DREADD recovers serum- and glucocorticoid-inducible kinase 2 (Sgk2) expression

  • •

    Sgk is required for chemogenetic rescue of social deficits and NMDAR hypofunction

Neuroscience; Behavioral Neuroscience; Molecular Neuroscience; Cellular Neuroscience

Introduction

Autism spectrum disorders (ASDs) are a group of neurodevelopmental disorders characterized by persistent deficits in social communication and interaction, as well as restricted, repetitive patterns of behavior, interests, or activities. At present, there is no US Food and Drug Administration-approved medical treatment for core symptoms of ASDs, which makes the discovery of novel interventions an important undertaking in ASD research (Lee et al., 2016).

One of the key brain regions impaired in ASDs is the prefrontal cortex (PFC), which mediates “high-level” executive functions, including working memory, sustained attention, decision making, and emotional control (Davidson, 2002, Arnsten and Rubia, 2012). Glutamatergic transmission is pivotal for PFC-dependent functions, and disturbed glutamatergic system in frontal cortex has been identified in patients with ASD (Stoner et al., 2014). One of ASD risk factors is the N-methyl-d-aspartate (NMDA)-type glutamate receptor (Hu et al., 2016, Rossi et al., 2017), which is required for neuronal survival, synaptic plasticity, and memory formation (Ikonomidou et al., 1999). NMDA receptor (NMDAR) blockade or NR1 deficiency induces ASD-like social deficits in mice (Zou et al., 2008, Billingslea et al., 2014, Wesseling et al., 2014).

Heterogeneous genetic variations are associated with ASD (Abrahams et al., 2013), whereas haploinsufficiency of the Shank3 gene has been identified as a prevalent and highly penetrant monogenic cause of autism (Durand et al., 2007, Betancur and Buxbaum, 2013). Shank3 encodes a scaffolding protein at postsynaptic density of glutamatergic synapses (Naisbitt et al., 1999, Hayashi et al., 2009). Shank3-deficient mouse models exhibit ASD-like phenotypes (Wang et al., 2011, Wang et al., 2016, Kouser et al., 2013, Duffney et al., 2015, Qin et al., 2018), as well as the significantly diminished NMDAR synaptic function in PFC (Duffney et al., 2015, Qin et al., 2018, Wang et al., 2019).

To identify the neuronal circuits and molecular targets that could be involved in ASD treatment, we used the chemogenetic tool, designer receptors exclusively activated by designer drugs (DREADDs), to manipulate the activity of PFC neurons in wild-type or Shank3-deficient mice. The Gq-coupled hM3Dq DREADD or Gi-coupled hM4Di DREADD, which is activated by exogenous administration of the small molecule clozapine-N-oxide (CNO), induces neuronal burst firing or neuronal silencing, respectively (Rogan and Roth, 2011). These reverse-engineered G protein-coupled receptors allow the remote, non-invasive, and long-lasting modulation of signal transduction pathways in discrete neuronal populations in vivo with therapeutic utility (Wess et al., 2013, Urban and Roth, 2015).

Results

Chemogenetic Activation of PFC Pyramidal Neurons Rescues Social Behavioral Deficits in Shank3-Deficient Mice

Shank3-deficient mice exhibited autism-like social preference deficits (Duffney et al., 2015, Qin et al., 2018). To determine whether this behavioral abnormality can be ameliorated by activating PFC pyramidal neurons, we injected the mCherry-tagged CaMKII-driven hM3D (Gq)-coupled DREADD adeno-associated virus (AAV) (Wei et al., 2016, Wang et al., 2018) bilaterally into the medial PFC (Figure 1A). Layer 5 pyramidal neurons, which showed the clearest deficits in autistic children (Stoner et al., 2014), were selected for electrophysiological recordings (Yuen et al., 2012, Zhong and Yan, 2016, Qin et al., 2018). As shown in Figures 1B and 1C, application of CNO (3 mg/kg, intraperitoneally [i.p.]) significantly increased the frequency of spontaneous action potentials (sAPs) in hM3D(Gq)-expressing PFC pyramidal neurons in both wild-type (WT) and Shank3-deficient mice (WT + Gq + saline: 1.34 ± 0.12 Hz, n = 22 cells/5 mice; WT + Gq + CNO: 3.59 ± 0.46 Hz, n = 15 cells/4 mice; Shank3+/ΔC + Gq + saline: 1.29 ± 0.17 Hz, n = 20 cells/5 mice; Shank3+/ΔC + Gq + CNO: 4.62 ± 0.23 Hz, n = 12 cells/4 mice; F1,65(treatment) = 118.7, p < 0.0001, two-way ANOVA), confirming the elevation of neuronal activity by this chemogenetic approach.

Figure 1.

Figure 1

Open in a new tab

Chemogenetic Activation of Pyramidal Neurons in PFC Ameliorates Social Behavioral Deficits in Shank3-Deficient Mice

(A) Low- and high-magnification images showing hM3Dq AAV expression in PFC pyramidal neurons.

(B) Quantification of sAP frequencies in PFC pyramidal neurons from WT versus Shank3+/ΔC mice (hM3Dq-injected) treated with saline or CNO (3 mg/kg, i.p.). ***p < 0.001, two-way ANOVA.

(C) Representative sAP traces.

(D) Bar graphs showing the time spent investigating either the social (Soc) or non-social (NS) stimulus in three-chamber sociability tests of young (5–6 weeks old) WT versus Shank3+/ΔC (hM3Dq-injected) mice treated with saline or CNO. +++p < 0.001, Soc versus NS; *p < 0.05, **p < 0.01, two-way ANOVA.

(E) Scatterplots showing the social preference index of the four groups. ***p < 0.001, two-way ANOVA.

(F) Representative heatmaps illustrating the time spent in different locations of the three chambers from the social preference tests.

(G) Bar graphs showing total sniffing time in social approach tests of the four groups. *p < 0.05, two-way ANOVA.

(H) Representative heatmaps illustrating the time spent in different locations of the apparatus from the social approach tests.

(I) Scatterplots of social preference index in Shank3+/ΔC mice (hM3Dq-injected) treated with CNO (two rounds) at different time points (1 h or 24 h after CNO). ***p < 0.001, ANOVA.

(J–L) Bar graphs showing a variety of behavioral tests of the four groups, including the latency to fall in rotarod tests (J), the time spent in the center in open-field tests (K), and the distance traveled in locomotion tests (L).

Next, we examined the impact of chemogenetic activation of PFC pyramidal neurons on autism-like social deficits in young (5–6 weeks old) male Shank3+/Δ mice. As shown in Figure 1D, in the three-chamber sociability assay, CNO (3 mg/kg, i.p.)-treated Shank3+/Δ mice (hM3Dq-injected) spent significantly more time exploring the social stimulus over the non-social object, whereas saline-treated Shank3+/ΔC mice showed a significant loss of the preference for the social stimulus and WT mice (hM3Dq-injected) treated with CNO had unchanged social preference (F3,88 (interaction) = 7.12, p < 0.001, n = 10–14 each group, two-way ANOVA). Social preference index was significantly elevated in Shank3+/ΔC mice after CNO treatment (Figures 1E and 1F, WT + Gq + saline: 49.1% ± 2.82%, n = 10 mice; WT + Gq + CNO: 51.7% ± 2.94%, n = 10 mice; Shank3+/ΔC + Gq + saline: 18.0% ± 3.43%, n = 14 mice; Shank3+/ΔC + Gq + CNO: 53.2% ± 3.50%, n = 14 mice, F1,44(treatment) = 31.37, p < 0.001, two-way ANOVA), indicating the dramatic alleviation of autism-like social preference deficits. To determine the specificity of the increased social preference by chemogenetic activation of PFC pyramidal neurons, we compared the preference of animals during phase I in the three-chamber sociability assay when they were exposed to two identical non-social stimuli (Duffney et al., 2015). CNO or saline-treated Shank3+/ΔC mice (hM3Dq-injected) showed almost no preference on these non-social objects (preference index, Shank3+/ΔC + Gq + saline: 4.4% ± 5.6%, n = 7 mice; Shank3+/ΔC + Gq + CNO: 5.6% ± 4.0%, n = 9 mice, p = 0.86, t test), indicating that the increase in preference seen in Shank3+/ΔC mice after DREADD excitation of the medial prefrontal cortex (mPFC) is specific for social stimuli. Although systemically injected CNO can convert to clozapine (Gomez et al., 2017), clozapine treatment (5 mg/kg, once daily for 3 days) itself failed to increase social interaction time or social preference in Shank3+/Δ mice (Qin et al., 2018).

We further examined social approach behaviors in Shank3+/ΔC mice with mPFC injection of hM3Dq. As shown in Figures 1G and 1H, Shank3+/ΔC mice treated with CNO spent significantly more time approaching and interacting with social stimuli than saline-treated Shank3+/ΔC mice, whereas CNO did not change social approach behaviors in WT mice (WT + Gq + saline: 128.8 ± 9.3 s; WT + Gq + CNO: 116 ± 12.1 s; Shank3+/ΔC + Gq + saline: 81.3 ± 10.8 s; Shank3+/ΔC + Gq + CNO: 127.7 ± 14.7 s, n = 13 each group, F1,48(interaction) = 6.22, p < 0.05, two-way ANOVA).

To find out the duration of the rescuing effect of chemogenetic activation of PFC pyramidal neurons, we performed social preference assays in Shank3+/ΔC mice (hM3Dq-injected) at various time points after CNO treatment. The significantly increased social preference index in Shank3+/ΔC mice was observed at 1 h, but not 24 h, after CNO injection, and re-injection of CNO significantly increased social preference index again, but the effect disappeared after 24 h (Figure 1I, WT + Gq + CNO: 56.7% ± 3.7%; Shank3+/ΔC + Gq + first CNO: 51.2% ± 5.3% [1 h], 23.4% ± 3.5% [24 h]; Shank3+/ΔC + Gq + second CNO: 56.7% ± 5.1% [1 h], 23.0% ± 4.6% [24 h]; n = 8 each group. F4,35 = 17.2, p < 0.0001, one-way ANOVA). These results suggest that chemogenetic activation of PFC pyramidal neurons is capable of rescuing social deficits in Shank3+/ΔC mice in a fast, short-lived, but repeatable manner.

A variety of other behaviors were also examined in Shank3+/ΔC mice with chemogenetic activation of PFC pyramidal neurons. As shown in Figures 1J–1L, no differences were observed in locomotion, rotarod, and open-field tests between WT and Shank3+/ΔC mice (hM3Dq-injected) treated with saline or CNO (n = 11–12 each group), suggesting that CNO does not alter motor coordination or anxiety-like behaviors.

Chemogenetic Activation of PFC Pyramidal Neurons Restores NMDAR Function and Facilitates AMPAR Functions in Shank3-Deficient Mice

Our previous studies found that NMDAR function in PFC was selectively diminished in Shank3-deficient mice (Duffney et al., 2015, Qin et al., 2018). Next, we measured the impact of chemogenetic activation of PFC pyramidal neurons on NMDAR-mediated excitatory postsynaptic current (EPSC) in WT and Shank3+/ΔC mice. We found that CNO treatment (1–2 h) restored the input/output curves of NMDAR-EPSC in PFC neurons of Shank3+/ΔC mice (hM3Dq-injected) and elevated NMDAR-EPSC in PFC neurons of WT mice (hM3Dq-injected) (Figure 2A, F3,44(treatment) = 23.6, p < 0.0001, n = 12 cells/3 mice for each group, two-way repeated measures ANOVA (rmANOVA)). No rescue of NMDAR-EPSC was found at 24 h after CNO injection (Figure S1A, F3,44(treatment) = 15.02, p < 0.0001, n = 12 cells/3 mice for each group, two-way rmANOVA).

Figure 2.

Figure 2

Open in a new tab

Chemogenetic Activation of PFC Pyramidal Neurons Restores NMDAR Function and Synaptic Expression in Shank3-Deficient Mice

(A) Summarized input-output curves of NMDAR-EPSC in layer V PFC pyramidal neurons from WT versus Shank3+/ΔC mice (hM3Dq-injected) treated with saline or CNO. *p < 0.05, **p < 0.01, ***p < 0.001, two-way rmANOVA.

(B) Confocal images (40×) of NR2A staining in PFC slices from the four groups.

(C) Quantification of the integrated intensity of NR2A puncta in the four groups. ***p < 0.001, **p < 0.01, two-way ANOVA.

To directly visualize synaptic NMDARs, we performed immunostaining of NR2A, a major subunit of synaptic NMDARs. As shown in Figures 2B and 2C, CNO-treated Shank3+/ΔC mice (hM3Dq-injected) had a significant increase in the expression level of NR2A puncta in PFC neurons, compared with saline-treated Shank3+/ΔC mice (integrated intensity: WT + Gq + saline: 100 ± 7.5; WT + Gq + CNO: 124.6 ± 4.0; Shank3+/ΔC + Gq + saline: 68.3 ± 4.2; Shank3+/Δ+Gq + CNO: 110 ± 4.4, F1,42(treatment) = 38.7, p < 0.0001, n = 10–12 images/3 mice each group). Co-staining of NR2A and PSD-95 (a synaptic marker) confirmed the synaptic location of NR2A puncta (Figure S2A). These data indicate that the diminished synaptic expression and function of NMDARs in Shank3-deficient mice are normalized by chemogenetic activation of PFC pyramidal neurons.

For AMPAR-EPSC, no significant difference was found between saline-treated WT and Shank3+/ΔC mice (hM3Dq-injected); however, CNO administration (1–2 h) significantly increased the amplitude of AMPAR-EPSC evoked by a series of stimulation intensities in both WT and Shank3+/ΔC mice (Figure 3A, F3,44(treatment) = 42.7, p < 0.0001, n = 12 cells/4 mice for each group, two-way rmANOVA). Spontaneous EPSC (sEPSC), a response resulting from spontaneously released glutamate, also had no difference in saline-treated WT and Shank3+/ΔC mice (hM3Dq-injected), but CNO treatment markedly increased sEPSC amplitude and frequency in both genotypes (Figures 3B and 3C, WT + Gq + saline: 12.56 ± 0.37 pA, 0.8 ± 0.09 Hz, n = 10 cells/3 mice; WT + Gq + CNO: 17.66 ± 1.84 pA, 1.61 ± 0.29 Hz, n = 8 cells/3 mice; Shank3+/ΔC + Gq + saline: 12.14 ± 0.29 pA, 0.75 ± 0.07 Hz, n = 10 cells/3 mice; Shank3+/ΔC + Gq + CNO: 17.13 ± 2.03 pA, 1.60 ± 0.3 Hz, n = 8 cells/3 mice, F1,32 (treatment) = 19.37, p = 0.0001, two-way ANOVA). No effect on AMPAR-EPSC was observed at 24 h after CNO injection (Figure S1B, F3,44 (treatment) = 0.38, p = 0.77, n = 12 cells/3 mice for each group, two-way rmANOVA).

Figure 3.

Figure 3

Open in a new tab

Chemogenetic Activation of PFC Pyramidal Neurons Facilitates AMPAR Function in Shank3-Deficient Mice

(A) Summarized input-output curves of AMPAR-EPSC in layer V PFC pyramidal neurons from WT versus Shank3+/ΔC mice (hM3Dq-injected) treated with saline or CNO. **p < 0.01, ***p < 0.001, two-way rmANOVA.

(B and C) Bar graphs of sEPSC amplitude (B) and frequency (C) in PFC pyramidal neurons from the four groups. *p < 0.05, two-way ANOVA. Inset: representative sEPSC traces.

(D) Confocal images (40×) of GluR2 staining in PFC slices from the four groups.

(E) Quantification of the integrated intensity of GluR2 puncta in the four groups. *p < 0.05, **p < 0.01, two-way ANOVA.

Immunostaining of GluR2 was also performed to visualize synaptic AMPARs. As shown in Figures 3D and 3E, despite the lack of difference in the expression level of GluR2 puncta in saline-treated WT and Shank3+/ΔC mice (hM3Dq-injected), CNO treatment caused a modest but significant increase of GluR2 expression in both WT and Shank3+/ΔC mice (integrated intensity: WT + Gq + saline: 100 ± 2.8; WT + Gq + CNO: 115.5 ± 3.4; Shank3+/ΔC + Gq + saline: 99.4 ± 5.5; Shank3+/ΔC + Gq + CNO: 123.8 ± 2.8, F1,42(treatment) = 29.7, p < 0.0001, n = 10–13 images/3 mice each group). The synaptic location of GluR2 puncta was confirmed by co-staining of GluR2 with PSD-95 (Figure S2B). These results indicate that the synaptic expression and function of AMPARs are elevated by chemogenetic activation of PFC pyramidal neurons.

Sgk Is Required for Chemogenetic Rescue of Social and Synaptic Deficits in Shank3-Deficient Mice, as well as for Normal Synaptic Function and Social Behaviors in WT Mice

Next, we examined potential molecular mechanisms underlying the rescuing effect of chemogenetic activation of PFC pyramidal neurons in Shank3-deficient mice. Our previous studies have found that the synaptic delivery of NMDARs is impaired in PFC neurons of Shank3+/ΔC mice (Duffney et al., 2015, Qin et al., 2018), and the membrane trafficking of glutamate receptors is facilitated by serum- and glucocorticoid-inducible kinases (Sgk, Yuen et al., 2011). Sgk is a family of immediate-early genes and serine/threonine kinases, which plays an important role in the regulation of ion channels, enzyme activity, gene transcription, hormone release, and neuroexcitability (Lang et al., 2006). The onset kinetics of the CNO effect (>1 h) is consistent with the time frame of Sgk activation, suggesting the potential involvement of Sgk in the chemogenetic rescue.

We first compared the expression level of Sgk in WT and Shank3+/ΔC mice (hM3Dq injected) treated with saline or CNO (1–2 h). Other related molecules, including calcium- and calmodulin-dependent protein kinase II (CaMKII), Src family tyrosine kinase Fyn, mGluR- and Shank-interacting protein Homer1, and activity-regulated cytoskeleton-associated protein Arc, were also compared. Quantitative PCR analyses (Figure 4A) showed that Shank3 deficiency induced a significant decrease of the mRNA level of Sgk2, Homer1a, Homer1b, and Arc, but did not change the mRNA level of Sgk1, Sgk3, CaMKIIα, or Fyn. CNO treatment of Shank3+/ΔC mice restored the expression of Sgk2 (F1,24 (genotype) = 82.83, p < 0.0001, F1,24 (treatment) = 23.93, p < 0.0001, n = 6–8/each group, two-way ANOVA), but not Homer1a/b or Arc. The expression of these genes might require stronger or prolonged stimulation, and additional work would be needed to clarify the specific reason for the differences in expression of these genes versus Sgk2. CNO treatment of WT mice elevated Sgk2 and Sgk3 mRNA levels. The restoration of Sgk2 expression in Shank3+/ΔC mice was not observed at 24 h after CNO injection (Figure S3).

Figure 4.

Figure 4

Open in a new tab

Sgk Is Required for Chemogenetic Rescue of Social and Synaptic Deficits in Shank3-Deficient Mice, as well as for Normal Synaptic Function and Social Behaviors in WT Mice

(A) Quantitative PCR data on the mRNA level of Sgk1/2/3, CaMKIIα, Fyn, Homer1a/b, and Arc in WT versus Shank3+/ΔC mice (hM3Dq-injected) treated with saline or CNO (3 mg/kg, i.p.). *p < 0.05, **p < 0.01, ***p < 0.001.

(B) Bar graphs showing the time spent investigating social (Soc) or non-social (NS) stimulus in three-chamber sociability tests of Shank3+/ΔC mice (hM3Dq-injected) injected (i.v.) with TAT-Sgk peptide (50 pmol/g, i.v.) or TAT peptide before CNO treatment. +++p < 0.001, Soc versus NS; *p < 0.05, two-way ANOVA.

(C) Scatterplots showing the preference index of the sociability tests of the two groups. ***p < 0.001, t test.

(D) Summarized input-output curves of NMDAR-EPSC and representative traces in layer V PFC pyramidal neurons from the two groups, *p < 0.05, **p < 0.01, ***p < 0.001, two-way rmANOVA.

(E and F) Summarized input-output curves of NMDAR-EPSC (E) and AMPAR-EPSC (F) and representative traces in layer V PFC pyramidal neurons from WT mice treated with TAT-Sgk peptide or TAT peptide. *p < 0.05, **p < 0.01, ***p < 0.001, two-way rmANOVA.

(G) Graphs showing the time spent investigating either the Soc or NS stimulus in 3-chamber sociability tests of WT mice injected with TAT-Sgk peptide or TAT peptide. +p < 0.05, +++ p < 0.001, Soc versus NS; ***p < 0.001, two-way ANOVA.

(H) Scatterplots showing the preference index of the two groups. ***p < 0.001, t test. Inset: representative heatmaps of the 3-chamber social preference test.

To find out whether the selective restoration of Sgk2 expression is required for the behavioral and synaptic rescue by chemogenetic activation of PFC pyramidal neurons, we inhibited Sgk function in Shank3-deficient mice. Sgk phosphorylates serine and threonine residues in the motif R-X-R-X-X-(S/T), so we used an Sgk substrate peptide (RPRAATF), which competitively blocks the interaction of Sgk with its endogenous substrates. The peptide was coupled to the protein transduction domain of the human immunodeficiency virus (HIV) TAT protein (YGRKKRRQRRR), which rendered it cell and blood-brain barrier permeant. Our previous studies have shown that Sgk function in PFC neurons can be inhibited by the systemic injection (intravenous [i.v.]) of the TAT-Sgk peptide (Yuen et al., 2011).

Shank3+/ΔC mice (hM3Dq-injected) were injected with TAT-Sgk peptide (50 pmol/g, i.v.) or a control TAT peptide at 30 min before CNO treatment. From the three-chamber sociability assay, we found that in Shank3+/ΔC mice injected with TAT-Sgk peptide, CNO (3 mg/kg, i.p.) treatment failed to elevate the social interaction time (Figure 4B, TAT-Sgk peptide: 68.9 ± 6.2 s (Soc [social]), 50 ± 5.6 s (NS [non-social]); TAT-peptide, 101 ± 12 s (Soc), 28.3 ± 2.8 s (NS), n = 8/group; F1, 28 (interaction) = 12.8, p < 0.01, two-way ANOVA) and failed to restore the social preference index (Figure 4C, TAT-Sgk peptide: 16.2% ± 5.9%, n = 8, TAT peptide: 54.7% ± 4.0%, n = 8, p < 0.001, t test). In addition, electrophysiological studies found that CNO failed to restore NMDAR-EPSC in Shank3+/ΔC mice injected with TAT-Sgk peptide, but not TAT control peptide (Figure 4D, F1,22 (treatment) = 25.91, n = 12 cells/3 mice each group, p < 0.001, two-way rmANOVA). These data suggest that Sgk is required for the chemogenetic rescue of social deficits and NMDAR hypofunction in Shank3-deficient mice.

To further test the causal role of Sgk in synaptic and social deficits associated with autism, we inhibited Sgk function in WT mice and examined its impact on synaptic function and social behaviors. We found that TAT-Sgk peptide injection (50 pmol/g, i.v.) to WT mice caused a significant reduction of NMDAR-EPSC (Figure 4E, n = 16 cells/4 mice each group, F1,30 (treatment) = 26.53, p < 0.001, two-way rmANOVA) and AMPAR-EPSC (Figure 4F, n = 12 cells/4 mice each group, F1,22 (treatment) = 18.88, p < 0.001, two-way rmANOVA) in PFC pyramidal neurons. In the three-chamber sociability assay, WT mice injected with TAT-Sgk peptide exhibited significantly reduced social interaction time (Figure 4G, TAT peptide: 106 ± 8.1 s, TAT-Sgk peptide: 60.8 ± 3.4 s, n = 13 pairs; F1, 48(interaction) = 22.58, p < 0.001, two-way ANOVA) and social preference index (Figure 4H, TAT peptide: 50.1% ± 3.3%; TAT-Sgk peptide: 23.1% ± 3.0%. n = 13 pairs, p < 0.001, t test). Taken together, these results have provided evidence showing the necessity and sufficiency of Sgk in the regulation of glutamatergic synaptic functions and social behaviors.

Chemogenetic Inhibition of PFC Pyramidal Neurons in WT Mice Is Insufficient to Induce Autism-like Social Deficits and Change Sgk Expression

As chemogenetic activation of PFC with hM3D(Gq) rescued social deficits in Shank3-deficient mice, we further tested whether chemogenetic inactivation of PFC is able to induce autism-like social deficits in WT mice. The mCherry-tagged CaMKII-driven Gi-coupled hM4-muscarinic receptor (hM4Di AAV), which induces neuronal silencing upon its activation by CNO (Urban and Roth, 2015), was injected bilaterally into the medial PFC of WT mice. Application of CNO (3 mg/kg, i.p.) significantly decreased the frequency of sAP in PFC pyramidal neurons of hM4Di-injected WT mice (Figure 5A, WT + Gi + saline: 1.11 ± 0.09 Hz, n = 14 cells/4 mice; WT + Gi + CNO: 0.25 ± 0.04 Hz, n = 14 cells/4 mice; p < 0.0001, t test), confirming the inhibitory effect of this chemogenetic approach.

Figure 5.

Figure 5

Open in a new tab

Chemogenetic Inhibition of PFC Pyramidal Neurons in WT Mice Fails to Induce Autism-like Social Deficits and Changes Sgk Expression

(A) Quantification of sAP frequencies in PFC pyramidal neurons from WT mice (hM4Di-injected) treated with saline or CNO (3 mg/kg, i.p.). ***p < 0.001, t test. Inset: representative sAP traces.

(B) Bar graphs showing the time spent investigating social (Soc) or non-social (NS) stimulus in three-chamber sociability tests of WT mice (hM4Di-injected) treated with saline or CNO. +++p < 0.001, Soc versus NS, two-way ANOVA.

(C) Scatterplots showing the social preference index of the two groups.

(D) Bar graphs showing total sniffing time in social approach tests of the two groups.

(E–G) Bar graphs showing a variety of behaviors of the two groups, including the distance traveled in locomotion tests (E), the time spent in the center during open-field tests (F), and the latency to fall during rotarod tests (G).

(H) Quantitative PCR data on the mRNA level of Sgk1, Sgk2, and Sgk3 in PFC of the two groups.

(I and J) Summarized input-output curves of NMDAR-EPSC (I) or AMPAR-EPSC (J) in layer V PFC pyramidal neurons from WT mice (hM4Di-injected) treated with saline or CNO (3 mg/kg, i.p.).

In the three-chamber sociability test, WT mice (hM4Di-injected) treated with CNO (3 mg/kg, i.p) or saline had similar social interaction time (Figure 5B, n = 12 pairs) and social preference index (Figure 5C, p > 0.05, t test). Social approach behavior was unchanged by CNO treatment of WT mice (Figure 5D, n = 12 pairs). Locomotion, rotarod, and open-field tests also failed to find differences between CNO and saline-injected WT mice (Figures 5E–5G, n = 12 pairs). Moreover, quantitative PCR analyses showed no significant difference in the mRNA level of Sgk1, Sgk2, and Sgk3 between WT mice (hM4Di-injected) treated with CNO and saline (Figure 5H, n = 8 pairs). Next, we examined the impact of chemogenetic inactivation of PFC pyramidal neurons on glutamatergic transmission. No significant difference was found in WT mice (hM4Di-injected) treated with CNO (1–2 h) or saline on NMDAR-EPSC (Figure 5I, F1, 22 (treatment) = 0.09, p = 0.76, n = 12 cells/3 mice for each group, two-way rmANOVA) or AMPAR-EPSC (Figure 5J, F1, 22 (treatment) = 0.23, p = 0.64, n = 12 cells/3 mice for each group, two-way rmANOVA). These results suggest that dampening the PFC neuronal activity in WT mice is not sufficient to induce autism-like social deficits, which could be due to the lack of changes in Sgk expression and NMDAR function. These results are consistent with prior optogenetic studies that inhibiting mPFC does not disrupt normal social preference in WT mice (Yizhar et al., 2011).

Discussion

Using the DREADD-based strategy, which has the capability to precisely control long-range and local neural network (Alexander et al., 2009, Michaelides et al., 2013, Wess et al., 2013, Urban and Roth, 2015), we have found that chemogenetic activation of PFC pyramidal neurons rescues autism-like social deficits in Shank3-deficient mice. Consistently, cognitive and social impairments in another autism model, transgenic mice carrying 16p11.2 deletion (16p11+/−), are also ameliorated by elevating PFC neuronal activity with Gq-coupled DREADD (Wang et al., 2018). It suggests that manipulating key neuronal circuits could be an effective avenue to alleviate behavioral abnormalities of ASD.

The behavioral effects of chemogenetic activation of PFC pyramidal neurons in Shank3-deficient mice are associated with the restoration or facilitation of NMDAR and AMPAR synaptic function, consistent with the implication of glutamatergic deficits in ASD (Wang et al., 2011, Wang et al., 2016, Kouser et al., 2013, Duffney et al., 2015, Qin et al., 2018). The membrane trafficking of glutamate receptors is regulated by a wide variety of molecules, including scaffolding proteins, synaptic plasticity genes like Arc, and protein kinases like CaMKII, Fyn, and Sgk (Prybylowski et al., 2005, Chowdhury et al., 2006, Yuen et al., 2011). Dysregulated NMDAR trafficking and synaptic efficacy are suggested to contribute to neuropsychiatric disorders like schizophrenia and stress-related diseases (Lau and Zukin, 2007, Popoli et al., 2011).

In search of potential molecular targets that are involved in Shank3 deficiency-induced loss of synaptic NMDARs and chemogenetic elevation of NMDAR and AMPAR function, we have found that only the Sgk family member Sgk2 shows diminished expression in Shank3-deficient mice, which is rescued by Gq-DREADD. Sgk is a family of serine/threonine kinase that plays a key role in regulating the trafficking of ion channels and transporters (Lang et al., 2006). Our previous studies have shown that stimulation of Sgk transcription elevates the expression of surface NMDARs and AMPARs (Yuen et al., 2011) by facilitating Rab4-mediated recycling of glutamate receptors to the synaptic membrane (Liu et al., 2010). Sgk is also a family of immediate-early gene that can be transcriptionally activated by stress hormones and phosphoinositide-3 kinase (PI3K) signaling pathway (Park et al., 1999). Activation of protein kinase C (PKC) has been shown to potentiate the activity of PI3K (Ziemba et al., 2016). Therefore, Gq-DREADD may elevate the mRNA level of Sgk2 via the PKC-PI3K cascade. Our discovery that blocking Sgk function in Shank3-deficient mice prevents Gq-DREADD from restoring NMDAR function and rescuing social deficits, and blocking Sgk function in WT mice induces the loss of PFC glutamatergic transmission and autism-like social deficits, further confirms the necessity and sufficiency of Sgk in these processes. Results gained from this study have provided insights into the circuity intervention and molecular targets for ASD treatment.

Limitation of the Study

Our evidence has suggested that the restored Sgk2 expression in Shank3-deficient mice induced by chemogenetic activation of PFC may contribute to the restoration or facilitation of NMDAR and AMPAR synaptic function and the rescue of social deficits; however, indirect effects cannot be ruled out. Several immediate-early genes show activity-dependent expression, and this may be responsible for the increased expression of Sgk2. Additional work would be needed to confirm that activity-dependent expression is indeed responsible for upregulation of Sgk2. The other limitation is the lack of direct visualization of the synaptic expression of NMDARs and AMPARs elevated by chemogenetic activation of PFC pyramidal neurons. Co-staining NR2A or GluR2 with PSD-95 in Gq DREADD-infected brain slices in WT or Shank3-deficient mice would provide complementary evidence in support of the electrophysiological data, but cannot currently be performed due to technical limitations.

Methods

All methods can be found in the accompanying Transparent Methods supplemental file.

Acknowledgments

This work was supported by the Nancy Lurie Marks Family Foundation and grants from the National Institutes of Health (MH112237, MH108842, and DA037618 to Z.Y.; MH116453 to L.Q.). We also thank E.F. Trachtman and the Varanasi family for their donations.

Author Contributions

L.Q. performed electrophysiological, immunocytochemical, and biochemical experiments; analyzed data; and wrote the draft. K.M. performed behavioral experiments and analyzed data. Z.Y. supervised the experimental design and wrote the paper.

Declaration of Interests

The authors declare no competing financial interests.

Published: July 26, 2019

Footnotes

Supplemental Information can be found online at https://doi.org/10.1016/j.isci.2019.06.014.

Supplemental Information

Document S1. Transparent Methods and Figures S1–S3
mmc1.pdf (2.9MB, pdf)

References

  1. Abrahams B.S., Arking D.E., Campbell D.B., Mefford H.C., Morrow E.M., Weiss L.A., Menashe I., Wadkins T., Banerjee-Basu S., Packer A. SFARI Gene 2.0: a community-driven knowledgebase for the autism spectrum disorders (ASDs) Mol. Autism. 2013;4:36. doi: 10.1186/2040-2392-4-36. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Alexander G.M., Rogan S.C., Abbas A.I., Armbruster B.N., Pei Y., Allen J.A., Nonneman R.J., Hartmann J., Moy S.S., Nicolelis M.A. Remote control of neuronal activity in transgenic mice expressing evolved G protein-coupled receptors. Neuron. 2009;63:27–39. doi: 10.1016/j.neuron.2009.06.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Arnsten A.F., Rubia K. Neurobiological circuits regulating attention, cognitive control, motivation, and emotion: disruptions in neurodevelopmental psychiatric disorders. J. Am. Acad. Child Adolesc. Psychiatry. 2012;51:356–367. doi: 10.1016/j.jaac.2012.01.008. [DOI] [PubMed] [Google Scholar]
  4. Betancur C., Buxbaum J.D. SHANK3 haploinsufficiency: a “common” but underdiagnosed highly penetrant monogenic cause of autism spectrum disorders. Mol. Autism. 2013;4:17. doi: 10.1186/2040-2392-4-17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Billingslea E.N., Tatard-Leitman V.M., Anguiano J., Jutzeler C.R., Suh J., Saunders J.A., Morita S., Featherstone R.E., Ortinski P.I., Gandal M.J. Parvalbumin cell ablation of NMDA-R1 causes increased resting network excitability with associated social and self-care deficits. Neuropsychopharmacology. 2014;39:1603–1613. doi: 10.1038/npp.2014.7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Chowdhury S., Shepherd J.D., Okuno H., Lyford G., Petralia R.S., Plath N., Kuhl D., Huganir R.L., Worley P.F. Arc/Arg3.1 interacts with the endocytic machinery to regulate AMPA receptor trafficking. Neuron. 2006;52:445–459. doi: 10.1016/j.neuron.2006.08.033. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Davidson R.J. Anxiety and affective style: role of prefrontal cortex and amygdala. Biol. Psychiatry. 2002;51:68–80. doi: 10.1016/s0006-3223(01)01328-2. [DOI] [PubMed] [Google Scholar]
  8. Duffney L.J., Zhong P., Wei J., Matas E., Cheng J., Qin L., Ma K., Dietz D.M., Kajiwara Y., Buxbaum J.D. Autism-like deficits in Shank3-deficient mice are rescued by targeting actin regulators. Cell Rep. 2015;11:1400–1413. doi: 10.1016/j.celrep.2015.04.064. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Durand C.M., Betancur C., Boeckers T.M., Bockmann J., Chaste P., Fauchereau F., Nygren G., Rastam M., Gillberg I.C., Anckarsater H. Mutations in the gene encoding the synaptic scaffolding protein SHANK3 are associated with autism spectrum disorders. Nat. Genet. 2007;39:25–27. doi: 10.1038/ng1933. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Gomez J.L., Bonaventura J., Lesniak W., Mathews W.B., Sysa-Shah P., Rodriguez L.A., Ellis R.J., Richie C.T., Harvey B.K., Dannals R.F. Chemogenetics revealed: DREADD occupancy and activation via converted clozapine. Science. 2017;357:503–507. doi: 10.1126/science.aan2475. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Hayashi M.K., Tang C., Verpelli C., Narayanan R., Stearns M.H., Xu R.M., Li H., Sala C., Hayashi Y. The postsynaptic density proteins Homer and Shank form a polymeric network structure. Cell. 2009;137:159–171. doi: 10.1016/j.cell.2009.01.050. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Hu C., Chen W., Myers S.J., Yuan H., Traynelis S.F. Human GRIN2B variants in neurodevelopmental disorders. J. Pharmacol. Sci. 2016;132:115–121. doi: 10.1016/j.jphs.2016.10.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Ikonomidou C., Bosch F., Miksa M., Bittigau P., Vockler J., Dikranian K., Tenkova T.I., Stefovska V., Turski L., Olney J.W. Blockade of NMDA receptors and apoptotic neurodegeneration in the developing brain. Science. 1999;283:70–74. doi: 10.1126/science.283.5398.70. [DOI] [PubMed] [Google Scholar]
  14. Kouser M., Speed H.E., Dewey C.M., Reimers J.M., Widman A.J., Gupta N., Liu S., Jaramillo T.C., Bangash M., Xiao B. Loss of predominant Shank3 isoforms results in hippocampus-dependent impairments in behavior and synaptic transmission. J. Neurosci. 2013;33:18448–18468. doi: 10.1523/JNEUROSCI.3017-13.2013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Lang F., Bohmer C., Palmada M., Seebohm G., Strutz-Seebohm N., Vallon V. (Patho)physiological significance of the serum- and glucocorticoid-inducible kinase isoforms. Physiol. Rev. 2006;86:1151–1178. doi: 10.1152/physrev.00050.2005. [DOI] [PubMed] [Google Scholar]
  16. Lau C.G., Zukin R.S. NMDA receptor trafficking in synaptic plasticity and neuropsychiatric disorders. Nat. Rev. Neurosci. 2007;8:413–426. doi: 10.1038/nrn2153. [DOI] [PubMed] [Google Scholar]
  17. Lee E., Lee J., Kim E. Excitation/inhibition imbalance in animal models of autism spectrum disorders. Biol. Psychiatry. 2016;81:838–847. doi: 10.1016/j.biopsych.2016.05.011. [DOI] [PubMed] [Google Scholar]
  18. Liu W., Yuen E.Y., Yan Z. The stress hormone corticosterone increases synaptic alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors via serum- and glucocorticoid-inducible kinase (SGK) regulation of the GDI-Rab4 complex. J. Biol. Chem. 2010;285:6101–6108. doi: 10.1074/jbc.M109.050229. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Michaelides M., Anderson S.A., Ananth M., Smirnov D., Thanos P.K., Neumaier J.F., Wang G.J., Volkow N.D., Hurd Y.L. Whole-brain circuit dissection in free-moving animals reveals cell-specific mesocorticolimbic networks. J. Clin. Invest. 2013;123:5342–5350. doi: 10.1172/JCI72117. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Naisbitt S., Kim E., Tu J.C., Xiao B., Sala C., Valtschanoff J., Weinberg R.J., Worley P.F., Sheng M. Shank, a novel family of postsynaptic density proteins that binds to the NMDA receptor/PSD-95/GKAP complex and cortactin. Neuron. 1999;23:569–582. doi: 10.1016/s0896-6273(00)80809-0. [DOI] [PubMed] [Google Scholar]
  21. Park J., Leong M.L., Buse P., Maiyar A.C., Firestone G.L., Hemmings B.A. Serum and glucocorticoid-inducible kinase (SGK) is a target of the PI 3-kinase-stimulated signaling pathway. EMBO J. 1999;18:3024–3033. doi: 10.1093/emboj/18.11.3024. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Popoli M., Yan Z., McEwen B.S., Sanacora G. The stressed synapse: the impact of stress and glucocorticoids on glutamate transmission. Nat. Rev. Neurosci. 2011;13:22–37. doi: 10.1038/nrn3138. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Prybylowski K., Chang K., Sans N., Kan L., Vicini S., Wenthold R.J. The synaptic localization of NR2B-containing NMDA receptors is controlled by interactions with PDZ proteins and AP-2. Neuron. 2005;47:845–857. doi: 10.1016/j.neuron.2005.08.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Qin L., Ma K., Wang Z.J., Hu Z., Matas E., Wei J., Yan Z. Social deficits in Shank3-deficient mouse models of autism are rescued by histone deacetylase (HDAC) inhibition. Nat. Neurosci. 2018;21:564–575. doi: 10.1038/s41593-018-0110-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Rogan S.C., Roth B.L. Remote control of neuronal signaling. Pharmacol. Rev. 2011;63:291–315. doi: 10.1124/pr.110.003020. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Rossi M., Chatron N., Labalme A., Ville D., Carneiro M., Edery P., des Portes V., Lemke J.R., Sanlaville D., Lesca G. Novel homozygous missense variant of GRIN1 in two sibs with intellectual disability and autistic features without epilepsy. Eur. J. Hum. Genet. 2017;25:376–380. doi: 10.1038/ejhg.2016.163. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Stoner R., Chow M.L., Boyle M.P., Sunkin S.M., Mouton P.R., Roy S., Wynshaw-Boris A., Colamarino S.A., Lein E.S., Courchesne E. Patches of disorganization in the neocortex of children with autism. N. Engl. J. Med. 2014;370:1209–1219. doi: 10.1056/NEJMoa1307491. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Urban D.J., Roth B.L. DREADDs (designer receptors exclusively activated by designer drugs): chemogenetic tools with therapeutic utility. Annu. Rev. Pharmacol. Toxicol. 2015;55:399–417. doi: 10.1146/annurev-pharmtox-010814-124803. [DOI] [PubMed] [Google Scholar]
  29. Wang W., Rein B., Zhang F., Tan T., Zhong P., Qin L., Yan Z. Chemogenetic activation of prefrontal cortex rescues synaptic and behavioral deficits in a mouse model of 16p11.2 deletion syndrome. J. Neurosci. 2018;38:5939–5948. doi: 10.1523/JNEUROSCI.0149-18.2018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Wang X., McCoy P.A., Rodriguiz R.M., Pan Y., Je H.S., Roberts A.C., Kim C.J., Berrios J., Colvin J.S., Bousquet-Moore D. Synaptic dysfunction and abnormal behaviors in mice lacking major isoforms of Shank3. Mol. Genet. 2011;20:3093–3108. doi: 10.1093/hmg/ddr212. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Wang X., Bey A.L., Katz B.M., Badea A., Kim N., David L.K., Duffney L.J., Kumar S., Mague S.D., Hulbert S.W. Altered mGluR5-Homer scaffolds and corticostriatal connectivity in a Shank3 complete knockout model of autism. Nat. Commun. 2016;7:11459. doi: 10.1038/ncomms11459. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Wang Z.J., Zhong P., Ma K., Seo J.S., Yang F., Hu Z., Zhang F., Lin L., Wang J., Liu T. Amelioration of autism-like social deficits by targeting histone methyltransferases EHMT1/2 in Shank3-deficient mice. Mol. Psychiatry. 2019 doi: 10.1038/s41380-019-0351-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Wei J., Xiong Z., Lee J.B., Cheng J., Duffney L.J., Matas E., Yan Z. Histone modification of Nedd4 ubiquitin ligase controls the loss of AMPA receptors and cognitive impairment induced by repeated stress. J. Neurosci. 2016;36:2119–2130. doi: 10.1523/JNEUROSCI.3056-15.2016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Wess J., Nakajima K., Jain S. Novel designer receptors to probe GPCR signaling and physiology. Trends Pharmacol. Sci. 2013;34:385–392. doi: 10.1016/j.tips.2013.04.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Wesseling H., Guest P.C., Lee C.M., Wong E.H., Rahmoune H., Bahn S. Integrative proteomic analysis of the NMDA NR1 knockdown mouse model reveals effects on central and peripheral pathways associated with schizophrenia and autism spectrum disorders. Mol. Autism. 2014;5:38. doi: 10.1186/2040-2392-5-38. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Yizhar O., Fenno L.E., Prigge M., Schneider F., Davidson T.J., O'Shea D.J., Sohal V.S., Goshen I., Finkelstein J., Paz J.T. Neocortical excitation/inhibition balance in information processing and social dysfunction. Nature. 2011;477:171–178. doi: 10.1038/nature10360. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Yuen E.Y., Liu W., Karatsoreos I.N., Ren Y., Feng J., McEwen B.S., Yan Z. Mechanisms for acute stress-induced enhancement of glutamatergic transmission and working memory. Mol. Psychiatry. 2011;16:156–170. doi: 10.1038/mp.2010.50. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Yuen E.Y., Wei J., Liu W., Zhong P., Li X., Yan Z. Repeated stress causes cognitive impairment by suppressing glutamate receptor expression and function in prefrontal cortex. Neuron. 2012;73:962–977. doi: 10.1016/j.neuron.2011.12.033. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Ziemba B.P., Burke J.E., Masson G., Williams R.L., Falke J.J. Regulation of PI3K by PKC and MARCKS: single-molecule analysis of a reconstituted signaling pathway. Biophys. J. 2016;110:1811–1825. doi: 10.1016/j.bpj.2016.03.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Zou H., Zhang C., Xie Q., Zhang M., Shi J., Jin M., Yu L. Low dose MK-801 reduces social investigation in mice. Pharmacol. Biochem. Behav. 2008;90:753–757. doi: 10.1016/j.pbb.2008.06.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Zhong P., Yan Z. Distinct physiological effects of dopamine D4 receptors on prefrontal cortical pyramidal neurons and fast-spiking interneurons. Cereb. Cortex. 2016;21:62–70. doi: 10.1093/cercor/bhu190. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Document S1. Transparent Methods and Figures S1–S3
mmc1.pdf (2.9MB, pdf)

Articles from iScience are provided here courtesy of Elsevier

RESOURCES