Summary
Reduced BDNF and GABAergic inhibition co-occur in neuropsychiatric diseases, including major depression. Genetic rodent studies show a causal link, suggesting the presence of biological pathways that mediate this co-occurrence. Here we show that mice with reduced Bdnf (Bdnf+/-) have upregulated expression of sequestosome-1/p62, an autophagy-associated stress response protein, and reduced surface presentation of α5 subunit-containing GABAA receptor (α5-GABAAR) in prefrontal cortex (PFC) pyramidal neurons. Reducing p62 gene dosage restored α5-GABAAR surface expression and rescued the PFC-relevant behavioral deficits of Bdnf+/- mice, including cognitive inflexibility and sensorimotor gating deficits. Increasing p62 levels was sufficient to recreate the molecular and behavioral profiles of Bdnf+/- mice. Finally, human postmortem corticolimbic transcriptome analysis suggested reduced autophagic activity in depression. Collectively, the data reveal that autophagy regulation through control of p62 dosage may serve as a mechanism linking reduced BDNF signaling, GABAergic deficits, and psychopathology associated with PFC functional deficits across psychiatric disorders.
HIGHLIGHTS BDNF constitutively promotes autophagy in cortical pyramidal neurons
Reduced BDNF causes elevated autophagy-regulator p62 expression, leading to lower surface α5-GABAAR presentation
Increasing p62 levels mimics cognition-related behavioral deficits in Bdnf+/- mice
Altered postmortem corticolimbic gene expression suggests reduced autophagic activity in depression
INTRODUCTION
Cognitive impairment is associated with a range of psychiatric and neurological conditions (Bast et al., 2017; Knight and Baune, 2018), manifesting either as a core symptom of major mental illnesses (e.g., major depressive disorder [MDD], schizophrenia), as age-related decline of brain functions, or as a condition comorbid to neurodegenerative and other brain disorders. Pathobiological mechanisms underlying cognitive impairment include deficits in neural plasticity or synaptic functions (Negrón-Oyarzo et al., 2016); however, the diversity of molecular entities and multiple neurotransmitter systems implicated in synaptic function and neuroplasticity have made it difficult to pinpoint relevant central neurobiological events contributing to cognitive dysfunction, and to identify therapeutic targets that can effectively alleviate these symptoms.
Translational molecular studies have consistently reported lower expression levels of brain-derived neurotrophic factor (BDNF) in postmortem samples from subjects with MDD (Guilloux et al., 2012; Tripp et al., 2012), schizophrenia (Weickert et al., 2003; Pillai et al., 2010; Islam et al., 2017), and age-related cognitive decline (Calabrese et al., 2013; Oh et al., 2016). BDNF is a member of the neurotrophin family of growth factors (Chao, 2003) which plays a number of critical roles in the nervous system, including synaptogenesis, neurotransmission, learning, memory, and cognition (Chao et al., 2006; Lu et al., 2014). Animal models with reduced BDNF expression or activity showed disturbances of neurotransmission or neural plasticity, which likely underlie cognitive deficits observed in these models (Dincheva et al., 2012; Lu et al., 2014; Dincheva et al., 2016), consistent with a dimensional contribution of this pathway in cognitive symptoms across a range of psychiatric disorders.
BDNF primarily signals through binding to TrkB receptor and its co-receptor, p75/NTR, leading to activation or modulation of downstream signaling molecules (Chao, 2003). BDNF signaling is also regulated at the receptor level. Upon ubiquitination of TrkB and p75/NTR by TRAF6, an E3 ligase, the ubiquitin-binding adaptor protein, sequestosome-1/p62, is recruited to form a protein complex (TrkB/p75/TRAF6/Ubi/p62), which is then trafficked to an appropriate cellular compartment (e.g., the proteasome or lysosome for degradation, the endosome for internalization or recycling), leading to down- or up-regulation of BDNF signaling in a context-dependent manner (Sánchez-Sánchez and Arévalo, 2017). Consistent with this model, a recent study reported that TrkB is located to the autophagosome and that it can mediate retrograde transport of this organelle in neurons (Kononenko et al., 2017).
p62 is a critical adaptor protein that integrates multiple cellular processes, including growth factor signaling, ubiquitin/proteasomal system, and autophagy/lysosomal system. This occurs by interacting with signaling molecules, ubiquitinated proteins, and autophagy-related protein LC3 (Lippai and Lőw, 2014). We recently reported that p62 regulates levels of the neuronal cell surface expression of γ-aminobutyric acid (GABA)A receptors (GABAARs) (Sumitomo et al., 2018a). In that study we show that in the prefrontal cortex (PFC) of mice heterozygous for Ulk2, an autophagy-regulatory gene, p62 protein levels are elevated as a result of attenuated autophagy, leading to sequestration of GABAA receptor-associated protein (GABARAP) (Pankiv et al., 2007), an adaptor protein implicated in endocytic trafficking of GABAARs (Wang et al., 1999). This leads to selective downregulation of GABAARs on the surface of pyramidal neurons, thereby underlying cognition-related behavioral deficits observed in Ulk2+/- mice (Sumitomo et al., 2018a).
Besides BDNF, we and others have consistently demonstrated dysfunctions of GABAergic inhibitory neurotransmission in the corticolimbic circuitry of MDD (Guilloux et al., 2012; Northoff and Sibille, 2014; Fee et al., 2017), schizophrenia (Caballero and Tseng, 2016; Hoftman et al., 2017), and age-related cognitive decline (Porges et al., 2017), suggesting a critical contribution of this inhibitory pathway to the pathophysiology of mental illnesses. Mouse–human translational studies further suggest that GABAergic changes occur downstream of reduced BDNF signaling, specifically affecting dendritic-targeting GABAergic interneurons (Guilloux et al., 2012; Tripp et al., 2012). Using human postmortem samples, we have demonstrated a positive correlation between BDNF and GABAergic synaptic gene expression, and in mouse models we showed that blockage of global or dendritic BDNF signaling in the PFC leads to reduction in expression of GABAergic genes mediating dendritic synaptic function (Oh et al., 2016; Oh et al., 2018), providing a causal link between reduced BDNF signaling and deregulated GABAergic neurotransmission. GABA elicits its inhibitory neurotransmission through pentameric GABAARs containing multiple subunits with diverse functional and anatomical properties, including somatically or perisomatically-targeted subunits (e.g., α1, α2) and dendritically-targeted subunits (e.g., α5).
Summing up the evidence: (1) BDNF signaling is reduced in neuropsychiatric conditions, (2) markers of GABAergic function are significantly decreased in neuropsychiatric conditions and in mice with reduced BDNF signaling, (3) dendritic BDNF transcripts and dendritically-localized α5-GABAAR are specifically affected in these conditions, (4) α5-GABAARs contribute to cognitive processes, and (5) autophagy-related protein (p62) regulates BDNF signaling and GABAAR trafficking. Accordingly, we tested the hypothesis that autophagy-related mechanisms operate downstream of BDNF, lead to the regulation of GABAergic functions preferentially through α5-GABAAR, and underlie deficits of cognitive processes observed across multiple neuropsychiatric conditions. We first investigated a putative new mechanism by which BDNF may control cell surface presentation of α5-GABAAR via regulation of p62 expression levels. We next investigated a functional and causal link between BDNF signaling and α5-GABAAR surface expression by bi-directional manipulation of p62 expression levels in mice and evaluation of cognitive flexibility and sensorimotor gating. We chose these two behavioral endophenotypes since they are relevant to psychiatric manifestations (Kellendonk et al., 2009; Parnaudesu et al., 2013; Swerdlow et al., 2016) and are commonly observed in BDNF mutant mice (Manning et al., 2013; Parikh et al., 2016) and α5-GABAAR-deficient mice (Hauser et al., 2005; Engin et al., 2013). Finally, we analyzed the relevance of the autophagy regulatory pathway to neuropsychiatric manifestations, using gene expression profiles obtained in postmortem brains of MDD and control subjects.
RESULTS
BDNF regulates autophagy in cortical neurons
Autophagy is a specialized membrane trafficking machinery and a major cellular recycling system primarily responsible for degrading old proteins and damaged organelles in the lysosome, which ultimately contributes to the maintenance of cellular homeostasis (Mizushima and Komatsu, 2011). Recent studies demonstrated new roles for autophagy in higher-order brain functions, through synapse pruning (Tang et al., 2014) or GABAA receptor trafficking (Sumitomo et al., 2018a), suggesting that autophagy deficits represent a pathological mechanism underlying cognition-related behavioral deficits relevant to neuropsychiatric disorders, including autism (Tang et al., 2014) and schizophrenia (Sumitomo et al., 2018a, b).
As a first step to address the role of BDNF in neuronal autophagy, we prepared primary cortical neurons from transgenic mice expressing GFP-LC3, a fluorescent marker for the autophagosome (Mizushima et al., 2004), and cultured them for 16 days before treating them with BDNF (100 ng/ml) for 30 min. We observed an increase in number and size of GFP-LC3-positive punctate structures (Figure 1A), suggesting either de novo autophagosome formation due to autophagy induction, or the accumulation of LC3 due to attenuated autophagic degradation. To discriminate between these two possibilities, we performed an autophagy flux assay in primary cortical neurons (Mizushima et al., 2010) and showed that BDNF increased the autophagy flux, i.e., the difference between the amount of membrane-bound LC3 (i.e., LC3-II) seen in the presence versus the absence of lysosomal protease inhibitors, which reflects the amount of LC3 degraded through an autophagy-dependent process within the lysosome (Figure 1B). These results suggest that BDNF has an autophagy-inducing activity in cortical neurons.
We next tested whether BDNF could also affect the later phase of autophagy (i.e., maturation stage), where proteins and organelles are degraded in the acidophilic lysosome. This can be assessed by evaluating the acidity of the autophagosome/lysosomal system using LysoTracker (Mizushima et al., 2010). To facilitate simultaneous observation of autophagosome formation and maturation, primary cortical neurons prepared from wild-type (WT) mice were transfected with GFP-LC3 and then labeled with LysoTracker for the last 5 min of the culture period, 30 min after adding BDNF. BDNF markedly increased the extent of LysoTracker uptake by the soma (Figure 1C), indicating that BDNF promoted maturation of the autophagy/lysosomal system in neurons. Consistently, neurons with greater levels of LysoTracker uptake also exhibited increase in number and size of autophagosomes located in neurites (Figure 1C, arrows), suggesting a sequence of events elicited by BDNF, from autophagosome formation to maturation. Together these data show that BDNF has an autophagy-enhancing activity in cultured cortical neurons.
Reduced BDNF expression leads to elevated p62 levels in cortical pyramidal neurons
To address the endogenous activity of BDNF in the regulation of autophagy in vivo, we quantitated the level of p62 protein in the medial prefrontal cortex (mPFC) of mice with reduced Bdnf levels (Bdnf+/- mice). In this brain region, BDNF is predominantly produced by CaMKII-positive pyramidal neurons and functions as an autocrine and paracrine factor to modulate the activity of neighboring excitatory and inhibitory neurons (Gorba and Wahle, 1999). p62 is used as an in vivo marker of autophagic activity, because it is the obligatory adaptor protein selectively targeted for autophagic degradation (Mizushima et al., 2010). The results show a significant increase in p62 protein levels in CaMKII-positive neurons of Bdnf+/- mice, as compared to WT control mice (Figure 1D and 1E). This increase in protein level was not paralleled by an increase in transcriptional activity of the p62 gene (Figure 1F), implying that the increased levels of p62 may result from reduced rates of protein degradation. These results suggest decreased autophagic activity in the presence of reduced BDNF levels, consistent with the hypothesis that BDNF constitutively enhances autophagy. To obtain further evidence in support of reduced autophagy in Bdnf+/- mice, we quantitated levels of expression of additional genes in this pathway. Expression of several autophagy regulatory genes (e.g., Lc3, Gabarap, Ulk2) remained unchanged, whereas expression of Ulk1, a gene critical to autophagy induction (Mizushima and Komatsu, 2011), was significantly downregulated by ~25% (Figure 1F). Together the data demonstrate a constitutive role for BDNF in enhancing autophagy, manifested by attenuated autophagy and persistent upregulation of p62 protein expression in the PFC of Bdnf+/- mice.
Elevated p62 expression in Bdnf+/- cortical neurons causes downregulated surface presentation of α5-GABAAR
We recently reported that, similar to Bdnf+/- mice, p62 protein expression levels are elevated in pyramidal neurons of the PFC in genetically-engineered mice with attenuated autophagy (Ulk2+/-), leading to downregulation of neuronal surface expression of GABAA receptors through sequestration of GABARAP, an adaptor protein responsible for endocytic trafficking of GABAA receptors (Sumitomo et al., 2018a). Besides, we previously showed that reduced BDNF signaling in the PFC causes decreased expression of GABA synaptic genes, most notably α5-GABAAR (Oh et al., 2016), a GABAA receptor subtype specifically expressed in the dendrites of pyramidal neurons (Fritschy and Mohler, 1995). These results suggest that BDNF affects GABA neurotransmission through regulation of α5-GABAAR levels in the dendrites of pyramidal neurons, consistent with Bdnf+/- mice exhibiting reduced amplitude and frequency of inhibitory miniature currents (mIPSC) in cortical and thalamic neurons (Laudes et al., 2012). We therefore reasoned that elevated p62 protein levels in Bdnf+/- neurons may influence the surface presentation of α5-GABAAR.
To test this hypothesis, we performed surface biotinylation of cultured cortical neurons and showed that cell surface α5-GABAAR protein levels were reduced in Bdnf+/- neurons, as compared with WT control neurons, with no significant changes in total levels of α5-GABAAR, or in both total and surface levels of the glutamate receptor NR1 (Figure 2A). Notably, reducing the p62 gene dosage in Bdnf1+/- neurons, using cortical neurons from Bdnf+/-;p62+/- mice, restored α5-GABAAR surface levels to WT levels (Figure 2A), suggesting that p62 is a critical adaptor mediating BDNF-induced altered surface presentation (i.e., trafficking) of α5-GABAAR.
To validate this finding in vivo, we performed receptor cross-linking assays using bis(sulfosuccinimidyl)suberate (BS3), a membrane-impermeable chemical cross-linker (Boudreau et al., 2012). As the cross-linking reaction proceeds in the presence of BS3, only the fraction of receptors expressed on the plasma membrane surface are expected to be covalently cross-linked with anonymous cell surface proteins, thereby transforming into higher molecular weight species, while the rest of the receptors associated with the endomembrane would remain intact, maintaining their original molecular weights. The PFC from WT and Bdnf+/- mice were subjected to the cross-linking reaction and the levels of a series of GABAAR subunits (i.e., α1, α2, and α5) were measured as a function of time (1–3 h). The results showed equivalent surface levels of α1- or α2-GABAAR (Figure S1), and a significant time-dependent decrease in the levels of intact α5-GABAAR (~50 kDa) in Bdnf+/- mice; specifically ~53% of total α5-GABAAR were estimated to be expressed on the cell surface in WT PFC (Figure 2B, left), whereas a lesser extent (~25%) of α5-GABAAR were presented on the cell surface under conditions of reduced Bdnf+/- levels (Figure 2B, middle). In contrast, ~56% of α5-GABAAR were observed in the PFC of Bdnf+/- in which p62 levels were genetically reduced (Bdnf+/-;p62+/-) (Figure 2B, right), demonstrating that reducing the p62 gene dosage restored the surface expression of α5-GABAAR to a level equivalent to WT mice. Collectively, the data demonstrate that decreased BDNF expression results in specific reduction in surface presentation of α5-GABAAR through elevated p62 expression in the PFC.
Elevated p62 expression in Bdnf+/- mice mediates behavioral deficits relevant to PFC dysfunction
We then investigated the potential role of elevated p62 expression in PFC-relevant brain functions of Bdnf+/- mice, such as information processing and cognition. Previous studies reported that Bdnf+/ mice have reduced prepulse inhibition (PPI) of acoustic startle response (Manning et al., 2013), demonstrating a role of BDNF in sensorimotor gating function. This mechanism of filtering sensory information to render appropriate motor responses has been shown to rely in part on the function of the cortical circuitry involving mPFC (Swerdlow et al., 2016). In agreement with the previous study, we first confirmed normal startle response and reduced PPI levels in Bdnf+/- mice (Figure 2C). We next showed that reducing the p62 gene dosage in Bdnf+/- mice (i.e., using Bdnf+/-;p62+/- mice) rescued the PPI deficits back to levels observed in control WT mice (Figure 2C). Together these results suggest that decreased expression of BDNF leads to sensorimotor gating deficits through elevated p62 expression.
Bdnf+/- mice were recently shown to exhibit reduced cognitive flexibility in a visual discrimination task (Parikh et al., 2016). To further address deficits in cognitive flexibility in Bdnf+/- mice, we used a rule shifting paradigm (Bissonette et al., 2008; Cho et al., 2015), in which mice were initially trained to associate food reward with a specific stimulus (i.e., either an odor or a digging medium) and subsequently evaluated for cognitive flexibility by changing the type of stimulus that predicts the reward (Figure 2D). Bdnf+/- and WT mice learned the association rule in a similar number of trials during the initial association phase of trials; however, Bdnf+/- mice required significantly higher numbers of trials to shift their behavior during the rule shifting phase of trials (Figure 2E). When the p62 gene dosage was reduced in Bdnf+/- mice (i.e., in Bdnf+/-;p62+/-), the mice showed cognitive performance indistinguishable from that of controls (Figure 2E), together suggesting that decreased expression of BDNF results in cognitive deficits through elevated p62 expression.
Elevated p62 expression is sufficient to cause downregulation of surface α5-GABAAR expression and behavioral deficits
To further address the causal role of elevated p62 expression in the regulation of α5-GABAAR surface expression and the associated behavioral changes, we generated p62-transgenic (Tg) mice, in which p62 transgene expression was driven by the CaMKII promoter. Among three Tg lines established, two lines (#1, #3) showed ~60% increase in p62 protein expression in the PFC, while one line (#2) failed to overexpress p62, as evaluated by Western blot (Figure 3A). BS3 cross-linking assays demonstrated reduced levels of surface α5-GABAAR expression in the PFC of the overexpressing p62-Tg line (#1) compared to WT, whereas the non-overexpressing line #2 displayed surface α5-GABAAR expression equivalent to WT levels (Figure 3B).
We next evaluated sensorimotor gating in the p62-Tg lines. p62-Tg lines #1 and 3 exhibited reduced levels of PPI, whereas the non-overexpressing line #2 had PPI levels that were not different from WT levels (Figure 3C). Similarly, in the cognitive flexibility test, mice from the overexpressing p62-Tg line (#1, #3) learned the association rule during the initial association phase of trials in a similar manner as WT, but were impaired during the rule shifting phase of trials (Figure 3D).
Together these data demonstrated that elevated p62 levels in CaMKII-expressing pyramidal neurons in the PFC and corticolimbic areas are sufficient to replicate the molecular and behavioral phenotypes of Bdnf+/- mice; specifically, reduced surface expression of α5-GABAAR and PFC-relevant behavioral deficits.
Human postmortem gene expression profiles suggests altered autophagy in depression
Although reduced levels of expression and/or functioning of BDNF and α5-GABAAR are consistently reported in the brains of several neuropsychiatric conditions, including MDD (Guilloux et al., 2012; Fee et al., 2017), the cellular machinery that may underlie these changes remains to be elucidated. On the basis of our recent reports showing attenuated neuronal autophagy in several mouse models for neuropsychiatric disorders such as schizophrenia (Sumitomo et al., 2018a, b), as well as the present data demonstrating elevated p62 levels in Bdnf+/- mouse model, we hypothesized that alteration in autophagy machinery may contribute to cellular deficits present in MDD.
To test this hypothesis, we analyzed genome-wide differential expression statistics from a meta-analysis of eight large-scale expression datasets in corticolimbic areas of 51 MDD patients and 50 controls (Ding et al., 2015). Autophagy-related genes were defined by gene ontology (GO) (Huang et al., 2009) and combined into “autophagy-enhancing” and “autophagy-attenuating” gene lists (See details in Tables S1 and S2). Of the 95 genes included in the “autophagy-enhancing” gene list, the expression of one gene was significantly increased (p<0.05), and 8 were significantly decreased (p<0.05) in MDD compared to controls (Table S1). By contrast, of the 38 genes included in the “autophagy-attenuating” gene list, the expression of 3 genes showed a significant increase and 2 genes a significant decrease in MDD compared to the control cohorts (Table S2). At the group level, an area under the curve (AUC) analysis revealed a significant difference (p<0.01) (Figure 4A), namely between an over-representation of upregulated autophagy-attenuating genes (AUC=0.60; 0.5 meaning no change) and a slight non-significant under-representation of downregulated autophagy-enhancing genes (AUC=0.44) in the MDD brain (Figure 4B). An alternate analysis using ranking of gene changes (gene set enrichment analysis, GSEA) of the same data showed no significant enrichment for autophagy-attenuating genes (Figure 4C, red line), but a significant enrichment in downregulation for autophagy-enhancing genes compared to controls (Figure 4C, green line, p=0.008). Together these results suggest reduced autophagy at the transcriptome level in MDD.
To further investigate autophagy-related gene expression profiles in different cohorts of MDD subjects, we performed RNAseq in the subgenual anterior cingulate cortex (sgACC) (Brodmann area 25) of postmortem samples obtained from four cohorts at different states of MDD (1: single episode, 2: first remission, 3: recurrent episode, and 4: second remission) and one control cohort (n=15–20/group; Figure S2) as described (Scifo et al., 2018). Applying GSEA to the data replicated a significant enrichment in upregulation for autophagy-attenuating genes in the combined MDD cohorts regardless of episode/remission status (1+2+3+4) compared to controls (Figure 4D, red line; Figure 4E, p=0.026). This enrichment in upregulation for the autophagy-attenuating gene set was also significant in remitted MDD subjects (cohorts 2+4) (Figure S3, red line; Figure 4E, p=0.016), but only at the trend level in currently-depressed subjects (Figure S4, red line; Figure 4E, p=0.098), suggesting either state-specific changes or reduced analytical power in smaller cohorts. No enrichment was observed for the autophagy-enhancing gene set in the combined or separate cohorts (Figure 4D, Figures S3 and S4, green line; Figure 4E).
Collectively, the GSEA results on the distinct cohorts and platforms (microarray and RNAseq datasets) are in good overall agreement and suggest persistent autophagy attenuation in MDD.
DISCUSSION
Building on our previous study showing a causal link between reduced BDNF signaling and selective attenuation of GABAergic gene expression in psychiatric disorders, such as MDD (Guilloux et al., 2012; Tripp et al., 2012; Oh et al., 2018) or during aging (Oh et al., 2016), we now demonstrate a novel mechanism by which (1) BDNF regulates autophagy in PFC pyramidal neurons and (2) reduced BDNF signaling negatively impacts GABA functions via autophagy-related control of GABAA receptor trafficking, and show that (3) these changes underlie behavioral manifestations that are relevant to PFC-mediated symptoms of psychiatric disorders. These results show for the first time that control of p62 levels, a molecule implicated in autophagic control of cellular function, is a key molecular event linking BDNF signaling, GABAergic neurotransmission, and specific behavioral manifestations related to PFC functions. Collectively, the results suggest the presence of coordinated biological processes linking BDNF to the maintenance of neuroplasticity and excitation–inhibition balance in the PFC. As we further show that gene implicated in attenuating autophagy tend to be upregulated in MDD, reduced autophagy may contribute to the pathology of the illness and mediate reduced GABAergic function downstream of reduced BDNF signaling.
The current study identifies p62, an autophagy regulator of cellular homeostatic processes, as a link between BDNF signaling and GABAergic function. p62 was originally identified as a protein induced by cellular stress conditions, and subsequently shown to function as an adaptor protein that integrates multiple cellular processes, including the autophagy/lysosomal pathway (Lippai and Lőw, 2014). Autophagy is activated in response to a range of cellular stresses (e.g., depletion of nutrients and energy, misfolded protein accumulation, oxidative stress) and mitigates such stresses to maintain cellular homeostasis (Mizushima and Komatsu, 2011). Using genetic mouse and cell models, the present study, as well as our recent report (Sumitomo et al., 2018a), demonstrated that autophagy can also control cell-to-cell signaling through regulation of surface expression of receptors mediating chemical inhibition. Thus, autophagy has the potential to serve as a homeostatic cellular machinery in the context of biological disturbances associated with brain disorders. Indeed, the gene expression data obtained from the brains of human subjects suggests that the overall activity of the autophagy machinery is attenuated in MDD, and the genetic rodent model studies suggest that elevated p62 expression may partly mediate the link between reduced BDNF signaling and GABAergic disruption through reduced GABAA receptor trafficking.
Reduced BDNF expression and deregulated GABA transmission frequently co-occur with psychiatric disorders, including MDD (Guilloux et al., 2012) and schizophrenia (Lewis et al., 2005), and during normal aging (Oh et al., 2016), suggesting a shared biological mechanism across these conditions. We previously demonstrated that reduced activity of BDNF, predominantly produced by pyramidal neurons, leads to reduced expression levels of presynaptic genes (e.g., Gad1, SLC32A1) and of neuropeptide genes (e.g., SST, neuropeptide Y, cortistatin) expressed in neighboring GABAergic inhibitory neurons targeting pyramidal neuron dendrites (Oh et al., 2016), together suggesting a paracrine mode of BDNF action responsible for attenuated GABA signaling (Figure 5, right). Moreover, expression levels of Gabra5, a gene encoding α5-GABAAR that is predominantly localized to the dendritic compartment of pyramidal neurons, were among the most significantly downregulated, suggesting an autocrine mode of BDNF action contributing to reduced expression of GABA-related genes. Both modes of transcriptional mechanisms, coupled with the attenuated GABAA receptor trafficking through the autophagy regulator, as demonstrated in the current study (Figure 5, left), are expected to synergistically reduce GABA neurotransmission across pre- and post-synaptic compartments.
The functional link between reduced BDNF signaling and GABA deficits across psychiatric disorders also suggests it may contribute to common behavioral endophenotypes across disorders. Cognitive function is primarily controlled via corticolimbic mechanisms, in part through regulation of excitatory and inhibitory neurotransmission (Kellendonk et al., 2009; Caballero and Tseng, 2016), and a number of synaptic modulators have been implicated in these processes, including BDNF and α5-GABAAR. Notably, reduced expression or activity of BDNF or α5-GABAAR in mice commonly manifests as cognitive inflexibility and sensorimotor gating deficits (Hauser et al., 2005; Engin et al., 2013; Manning and van den Buuse, 2013; Parikh et al., 2016). This suggests these molecules represent a network of regulators underlying cognitive processes commonly affected across psychiatric conditions.
There are several limitations to this study. Given the multiple roles and binding partners of p62 adaptor protein, it is unlikely that α5-GABAAR is the only receptor system or cellular target affected by elevated p62 levels, and additional cellular machineries may further contribute to the behavioral deficits in Bdnf+/- mice. Nonetheless we found here and in our prior transcriptomic studies that α5-GABAARs are preferentially affected, as opposed to α1-GABAARs or α2-GABAARs for instance, which are localized to somatic and perisomatic cellular compartments. Based on the predominant expression of BDNF in pyramidal neurons in the PFC and on the restricted expression of α5-GABAAR in corticolimbic brain regions, we primarily focused our analysis on the PFC-related molecular and behavioral deficits in the current study. Additional cell type- and brain region-specific approaches will be necessary to further determine the precise role of these molecules. Finally, these studies were performed in male mice and comparative analyses in female mice are warranted.
Nonetheless, given the critical role of p62 in regulating α5-GABAAR trafficking and behavioral outcomes in Bdnf+/- mice, we propose that these molecular players (i.e., p62, BDNF, α5-GABAAR) critically contribute to cognitive processes and other brain functions under both normal and pathophysiological conditions. For instance, p62 protein levels typically increase with age, reflecting a gradual decrease in cellular autophagic activity (Vilchez et al., 2014), and elevated p62 levels or increase in p62+ inclusions are cardinal features of neurodegenerative disorders, including Alzheimer’s and Parkinson’s diseases (Ferrer et al., 2011; Salminen et al., 2012). Furthermore, we recently reported upregulated expression of p62 proteins in cultured neurons isolated from subjects with schizophrenia and bipolar disorder (Sumitomo et al., 2018a), and in brains of a mouse model of schizophrenia (Sumitomo et al. 2018b). Hence, controlling the dosage of p62 protein may provide a potential target for therapeutic intervention against symptoms shared across these disorders, such as cognitive impairment, through augmentation of inhibitory neurotransmission. Of note, cognitive impairment is among symptoms most difficult to treat and is known to frequently persist during remission in MDD (Disner et al., 2011). The gene expression profiles in MDD in our present study therefore endorse the idea that attenuated autophagy throughout episode/remission states may provide a biological underpinning for persistent cognitive impairment in MDD. In addition, surface availability of GABAA receptors represents a rate-limiting step for GABAergic neurotransmission. Therefore, the mechanisms for regulating surface presentation of GABAA receptor through p62 dosage control may provide an alternative therapeutic approach, for instance for MDD subjects who do not respond to current antidepressant treatment, or for targeting cognitive deficits during remission of depression, across brain disorders and during aging.
MATERIALS AND METHODS
Animals
Bdnf knockout mice (Bdnftm1Jae/J, stock No. 002266) were obtained from the Jackson Laboratory (Bar Harbor, USA). GFP-LC3 transgenic mice were provided by Dr. Noboru Mizushima (University of Tokyo, Japan). p62 knockout mice were provided by Dr. Toru Yanagawa (University of Tsukuba, Japan). CaMKII-p62 transgenic mice were generated according to the standard procedures (Nagy et al., 2003). Mice were maintained on the C57BL/6J genetic background for at least 10 generations. Eight to 12-week old male mice were used for behavioral analysis. Behavioral experiments and data collection were performed by experimenters blind to animal genotypes. Maintenance of mouse colonies and experiments using mice were in accordance with the NIH Guide for the Care and Use of Laboratory Animals, and approved by the Institutional Animal Care and Use Committee at Kyoto University and at the Centre for Addiction and Mental Health.
Prepulse inhibition
The startle response and prepulse inhibition (PPI) were measured using a startle reflex measurement system (SR-LAB) as described (Sumitomo et al., 2018a). The test session began by placing a mouse in a plastic cylinder and leaving it undisturbed for 30 min. The background white noise level in the chamber was 70 dB. A prepulse–pulse trial started with a 50-ms null period, followed by a 20-ms prepulse white noise (74, 78, 82, 86, or 90 dB). After a 100-ms delay, the startle stimulus (a 40-ms, 120 dB white noise) was presented, followed by a 290-ms recording time. The total duration of each trial was 500 ms. A test session consisted of six trial types (pulse-only trial, and five types of prepulse–pulse trial). Six blocks of the six trial types were presented in a pseudo–randomized order such that each trial type was presented once within a block. The formula: 100– ([Response on acoustic prepulse-pulse stimulus trials/Startle response on pulse-only trials] x 100) was used to calculate %PPI.
Rule shift assay
Cognitive flexibility was evaluated in the rule shift assay, essentially as described (Bissonette et al., 2008; Cho et al., 2015). In brief, mice were habituated to food, feeding apparatus, different odor cues and digging medium texture cues prior to testing, and then food-deprived a day before the assays. Mice were initially trained in a sequence of trials to associate a food reward with a specific stimulus (i.e., either an odor or a digging medium). A varying combination of stimulus and food reward was presented to mice per trial. Eight consecutive correct responses to the food reward were considered reaching criterion (i.e., successful establishment of association between the stimulus and the food reward), and the number of trials to reach criterion were scored for each mouse before and after rule shifting (e.g., from an odor cue to a different texture cue predicting reward). Upon rule shifting, numbers of errors due to perseveration to an old rule were scored before reaching new criterion.
Quantitative reverse transcription-polymerase chain reaction (qRT-PCR)
Total RNA was extracted from the PFC using RNeasy Mini kit (Qiagen), and reverse-transcribed with a ReverTra Ace cDNA synthesis kit (Toyobo). TaqMan probes were purchased from Applied Biosystems, Inc. All data were normalized with Gapdh as reference.
Primary cortical neuron culture
Primary cortical neurons were prepared from E13.5 frontal cortex through papain treatment (0.5 μg/ml in Earle’s balanced salt solution supplemented with 5 mM EDTA and 200 μM L-cysteine), followed by mechanical trituration using fire-bore glass pipettes, and plated on poly-D-lysine-coated glass cover slips or glass-bottom dishes (MatTek). The cultures were recovered in serum-containing media (Neurobasal media supplemented with 10% horse serum, 5% fetal bovine serum, and 2 mM glutamine [Gibco]) for 4 h and maintained in serum-free media (Neurobasal media supplemented with B-27 (1:50 diluted), 2 mM glutamine, 50 I.U./ml penicillin, and 50 μg/ml streptomycin), with half of media being replaced with fresh media every 2–3 days. The cultures were used for immunofluorescence analysis or for surface biotinylation followed by Western blot analysis.
Neuronal autophagy and autophagy flux assays
For evaluation of autophagy, primary cortical neurons cultured for 18–25 days were transfected with GFP-LC3 expression plasmid using Lipofectamine 2000 (Invitrogen) via the standard procedure, and 48 h post-transfection, the cells were incubated with BDNF (Sigma, 100 ng/ml) for up to 40 min before fixation with 4% paraformaldehyde (PFA) in PBS. When appropriate, LysoTracker (Invitrogen) was included in culture media 5 min before fixation. Fluorescence images were acquired by confocal microscopy (SP8, Leica) and the fluorescence intensities of GFP+ puncta and LysoTracker uptake were evaluated by ImageJ (NIH). As for the autophagy flux assay (Mizushima et al., 2010), primary cortical neurons were treated with BDNF (100 ng/ml) in the presence or absence of a lysosomal protease inhibitor (bafilomycin A1 [BafA1], 5 ng/ml) and the cell lysates were analyzed by Western blot using LC3 antibody. The amounts of autophagosomal membrane-bound LC3 (LC3-II) normalized by total amounts of LC3 (cytosolic LC3-I plus membrane bound LC3-II) were compared in the presence versus the absence of BafA1 to calculate the autophagy flux, which corresponds to the amount of LC3 degraded through an autophagy-dependent process within the lysosome.
Surface biotinylation
Biotinylation of cell surface proteins was performed in primary neuron cultures using the cell surface protein isolation kit (Pierce) according to the manufacturer’s protocol. Briefly, cells were incubated with ice-cold PBS containing Sulfo-NHS-SS-Biotin (Pierce) for 30 min with gentle rocking at 4°C. Cells were then lysed and precipitated with NeutrAvidin beads. Precipitated proteins were eluted from the NeutrAvidin beads with loading buffer containing dithiothreitol (DTT) and heated for 5 min at 95°C and then analyzed by Western blot.
Chemical cross-linking assay
Cell surface receptor cross-linking assays were performed as described (Boudreau et al., 2012). In brief, mice were decapitated and coronal brain slices (~1 mm thick) were quickly prepared using Brain Matrix (Ted Pella) within 30 sec on ice. The PFC was then excised from the slices, minced into small pieces using a razor blade, and incubated in artificial CSF buffer containing bis(sulfosuccinimidyl)suberate (BS3) cross-linker (2 mM, ThermoFisher) for 30 min to 4 h at 4°C with constant invert mixing. After quenching the crosslinking reaction by adding glycine (100 mM) for 10 min at 4°C, the tissues were harvested by centrifugation (20,000 g, 4°C, 2 min). The proteins were prepared in lysis buffer containing 0.1 % Nonidet P-40 (v/v), protease and phosphatase inhibitor cocktail, and 1 mM DTT, and analyzed by Western blot.
Western blots
Western blot analysis was performed according to the standard procedure. The primary antibodies used were anti-α1-GABAA receptor (rabbit, 1:5,000, abcam), anti-α2-GABAA receptor (rabbit, 1:400, Alomone Labs), anti-α5-GABAA receptor (rabbit, 1:1,000, R&D Systems), anti-NR1 (rabbit monoclonal [1.17.2.6], 1:1,000, Millipore), anti-p62 (guinea pig, 1:1,000, MBL), anti-LC3B (rabbit, 1:1,000, Novus), anti-GAPDH (mouse, 1:1,000, abcam), and anti-α-Tubulin (mouse monoclonal [B-5-1-2], 1:8,000, Sigma).
Immunofluorescence
Brains of mice (n = 4 per group) perfused with 4% PFA/PBS were serially cut into 50 μm-thick coronal sections using vibratome (VT1200S, Leica), and the sections from one cohort of mice (a littermate pair of wild-type and Bdnf+/- mice) were permeabilized in PBS containing 0.05% Triton X-100 for 1 h, incubated for 30 min at room temperature in 10% goat serum (Chemicon) in PBS, and immunostained for 16 h at 4°C with primary antibodies followed by Alexa Fluor® 488- or 546-conjugated secondary antibodies (Molecular Probes) for 1 h at room temperature. The stained samples were observed using a confocal microscope (SP8, Leica; 40x objective lens, NA = 1.3); images of one optical section (1 μm thick) were acquired from 3 to 6 non-overlapping areas per section, randomly chosen in PFC (layer 2/3), and 3 to 4 serial sections were analyzed. The fluorescence intensities of p62 immunostaining were measured from each neuronal soma (30~50 somas per section) using ImageJ (NIH), with the background levels of staining in adjacent regions being subtracted, and the average immunofluorescence intensity was calculated across all serial sections from every mouse used. The primary antibodies used were: anti-p62 (guinea pig, 1:400, MBL) and anti-CaMKIIα (mouse monoclonal [6G9], 1:500, Stressmarq).
Human transcriptome analysis
For differential expression summary statistics, data from a prior meta-analysis of altered gene expression in MDD were used (Ding et al., 2015). In brief, human postmortem brain samples were obtained after consent from next of kin during autopsies conducted at the Allegheny County Medical Examiner’s Office (Pittsburg, PA, USA) using procedures approved by the Institutional Review Board and Committee for Oversight of Research Involving the Dead at the University of Pittsburgh. A total of 51 MDD and 50 control subjects were included in the 8 studies in that report. Samples from the dorsolateral prefrontal cortex (dlPFC), subgenual anterior cingulate cortex (sgACC) or rostral amygdala enriched in lateral, basolateral and basomedial nuclei had been previously collected and processed on Affymetrix HG-U133 Plus 2 or Illumina HT12 gene arrays. Four studies were performed in the sgACC, 2 studies in the amygdala and 2 in the dlPFC. Half of the studies had been performed in female subjects in each brain region. See details on subjects, areas investigated and other parameters in Ding et al., 2015. Differential summary statistics were used to rank the 10,621 genes from the upregulated gene with the lowest p-value to the downregulated gene with lowest p-value. The area under the receiver operating curve statistics was used to test enrichment of the autophagy-related gene lists. Significance between the area under the curve (AUC) values for the autophagy-enhancing and autophagy-attenuating gene lists was empirically determined using 10,000 comparisons of randomly selected gene sets of the same sizes.
For gene set enrichment analysis (GSEA) (Subramanian et al., 2005), RNAseq data on sgACC postmortem samples from four MDD cohorts (single episode, n = 20; first remission, n = 15; recurrent episode, n = 20; and second remission, n = 15) and one control cohort (n = 20) were used; sample collection procedures, the site of collection and approval body are essentially the same as above (see details in Scifo et al., 2018), and RNAseq procedure and bioinformatics analysis are as described (Shukla et al., 2018). Full datasets on RNAseq analysis will be reported elsewhere (Shukla et al., BioRxiv).
Statistical analysis
All data were represented as mean ± standard error of the mean (SEM) and were analyzed by Kruskal-Wallis test followed by Dunn’s multiple comparison test, unless otherwise noted. Behavioral assay data were analyzed by one-way or two-way analysis of variance (ANOVA) followed by Bonferroni post-hoc test, using Prism statistics software (GraphPad).
AUTHOR CONTRIBUTIONS
T.T. and E.S. conceived the studies; T.T., A.S., and Y.H.-T. carried out experiments; H.M. generated CaMKII-p62 transgenic mice; H.O. provided analytical tools; R.S. and L.F. analyzed gene expression profiles; T.T. and E.S. wrote and edited the paper.
DECLARATION OF INTERESTS
The authors declare no competing interests.
ACKNOWLEDGMENTS
This work was supported by grants from the Canadian Institute of Health Research (CIHR #153175 to E.S.), National Alliance for Research on Schizophrenia and Depression (NARSAD award #25637 to E.S.), the National Institutes of Health (MH-093723 to E.S.), Campbell Family Mental Health Research Institute (to E.S.), and Department of Defense/Congressionally Directed Medical Research Program (W81XWH-11-1-0269 to T.T.).