A developmental stage- and Kidins220-dependent switch in astrocyte responsiveness to brain-derived neurotrophic factor

Astroglia, the most abundant cell population of the central nervous system (CNS), are taking the stage in the neuroscience field because of their multifaceted role in the modulation of neural physiology at increasing levels of complexity, i.e. from the single synapse to higher circuits (Santello et al., 2019). Key to astrocyte performance is the capability to sense a wide range of extracellular signals and, consequently, activate intracellular signaling pathways. These can be kinase-based – as the ubiquitous mitogen-activated protein kinase (MAPK), e.g. SHP2, Ras, ERK, pathways – or comprising signal transducer and activator of transcription 1 and 3 (STAT1 and STAT3), or nuclear factor kappa B (NFκB) cascades (Colombo and Farina, 2016), or initiated via second messengers that trigger the release of Ca²⁺ ions from the endoplasmic reticulum (Okubo, 2020). Intracellular Ca²⁺ transients ([Ca²⁺]_i) underlie the vast majority of responses of astrocytes as they mediate the release of a plethora of soluble factors, i.e. gliotransmitters, that include neurotransmitters, trophic factors, hormones, cytokines and several other bio-active molecules. Gliotransmitters exert both autocrine and paracrine functions, thus controlling and coordinating the activity of astroglia and of neighboring neurons (Verkhratsky et al., 2016).

Although the mechanisms that underlie astrocyte uptake/release of glutamate, ATP and other gliotransmitters – as well as the ability to respond to neurotransmitters, such as glutamate – have been well-described and agreed upon by the scientific community, the extent to which astrocytes produce, sense and respond to neurotrophins, such as nerve-growth factor (NGF) and brain-derived neurotrophic factor (BDNF), is still an object of debate. Neurotrophins are mostly studied for their pleiotropic effects on neurons, as they control virtually every aspect of neuronal physiology, including maturation, axonal and dendritic differentiation, synaptogenesis and various forms of synaptic plasticity (Park and Poo, 2013). Besides their well-known effects on neurons, neurotrophins also modulate important aspects of astrocyte physiology (Poyhonen et al., 2019). Within the CNS, the most abundant neurotrophin is BDNF, which signals through its full-length receptor, the neurotrophic receptor tyrosine kinase 2 (TrkB, also known as NTRK2), or the kinase-deficient truncated version of it (TrkB-T1) (Park and Poo, 2013). Astrocytes predominantly express TrkB-T1, which triggers release of Ca²⁺ from intracellular stores (Rose et al., 2003) and activates various kinase pathways (Ohira et al., 2005; Vaz et al., 2011) independently of the full-length protein. The BDNF-TrkB-T1 system modulates several astrocyte properties, such as astrocyte morphological maturation, and their capability of sensing neurotransmission by controlling the membrane targeting of glycine and GABA transporters (Aroeira et al., 2015; Holt et al., 2019; Ohira et al., 2007; Vaz et al., 2011). An aberrant activation of this pathway has been described in the experimental autoimmune encephalomyelitis rodent model for multiple sclerosis, where overexpressed TrkB-T1 induces neuronal death by promoting the release of nitric oxide (Colombo et al., 2012). Moreover, astrocytic TrkB-T1 contributes to neuropathic pain and neurological dysfunctions in rodent models of spinal cord injury and of amyotrophic lateral sclerosis (Matyas et al., 2017; Yanpallewar et al., 2021). TrkB-T1 is, arguably, the predominant receptor isoform expressed in astrocytes; however, some reports have indicated that cultured astrocytes can also express full-length TrkB (Climent et al., 2000; Condorelli et al., 1995). Interestingly, some evidence indicates that TrkB expression by astrocytes occurs upon injury of the CNS or in the presence of chronic and inflammatory diseases, such as multiple sclerosis (McKeon et al., 1997; Soontornniyomkij et al., 1998; Stadelmann et al., 2002). Importantly, the competence of astrocytes to perceive and respond to BDNF seems to be different in various brain areas (Saba et al., 2020), thus adding a further layer of complexity to the system. Besides sensing extracellular BDNF through specific receptors, astrocytes also take up and recycle BDNF as well as its precursor form (proBDNF) in vitro and in vivo (Bergami et al., 2008; Vignoli et al., 2016). An increasing body of evidence suggests that cultured astrocytes produce and secrete BDNF under basal conditions (Kinboshi et al., 2017; Rousseaud et al., 2015), and that BDNF expression is increased by specific stimuli (Inoue et al., 1997; Jean et al., 2008; Miklic et al., 2004; Saha et al., 2006; Zafra et al., 1992). Albeit less studied than the BDNF-TrkB system, astrocytes also express the tumor necrosis factor receptor superfamily member 16 (NGFR, hereafter referred to as p75^NTR), which binds all pro-neurotrophins that mediate distinct signaling pathways with similar affinity. Astrocytes upregulate p75^NTR upon injuries or cellular insults, and astrocytic p75^NTR has been involved in the modulation of cell proliferation associated with astrocytosis and the formation of glial scars (reviewed by Cragnolini and Friedman, 2008).

Although the contribution of astrocytes to neuronal function is becoming clear, the cellular and molecular mechanisms that underlie the regulation of communication between astrocytes and neurons mediated by neurotrophins remains incompletely understood. Kinase D-interacting substrate of 220 kDa (Kidins220) is one of the several proteins involved in the neurotrophin pathways. Kidins220 is a four-pass transmembrane protein enriched in the nervous system, where it is expressed by both neurons and astrocytes, and involved in various aspects of neuronal differentiation and plasticity (Neubrand et al., 2012; Scholz-Starke and Cesca, 2016). Recently, we have shown that Kidins220 participates in the molecular events that induce spontaneous and evoked [Ca²⁺]_i transients in primary astrocytes, thereby controlling the vulnerability of astrocytes to genotoxic stress and the maturation of co-cultured neurons (Jaudon et al., 2020).

In this work, we investigated the responsivity of astrocytes to BDNF and the role of Kidins220 in astrocytic BDNF signaling. We first compared the signaling competence of embryonic and postnatal primary cortical astrocytes exposed to BDNF, and found a developmental shift from a kinase-based response in embryonic cells to a predominantly [Ca²⁺]_i-based response in postnatal cultures, which was paralleled by the reduced expression of full-length TrkB. We also addressed the role of Kidins220 in BDNF-induced signaling within astrocytes, and found that Kidins220 significantly contributes to both kinase and [Ca²⁺]_i pathways. Our data contribute to the understanding of the complex astrocytic response to neurotrophins, describing a developmental stage- and Kidins220-dependent switch in the responsiveness of astrocytes to BDNF.

To determine the role of Kidins220 in the modulation of BDNF signaling in astrocytes, we compared expression of TrkB in wild-type (+/+) and Kidins220^−/− primary astrocytes by using western blot analysis. Astrocytes were derived from E18.5 embryos and used after 15 days in vitro (DIV), a stage at which cultures had reached confluence. Compared with wild-type cells, Kidins220^−/− astrocytes displayed a significant reduction of both full-length and truncated TrkB forms (∼40% and 50%, respectively), although the ratio between these two isoforms remained unchanged (Fig. 1A). In contrast, no differences were detected in the expression of p75^NTR in the two genotypes (Fig. S1A). We then compared expression levels of the main downstream effectors of the BDNF-TrkB pathway, i.e. of mitogen-activated protein kinases 1 and 2 (MAPK1 and 2), phospholipase C-gamma-1 (PLCG1, hereafter referred to as PLCγ) and Akt. Whereas total protein levels of MAPK1/2 and Akt1, 2 and 3 (hereafter referred to as Akt) were unaffected, we found a significant reduction of PLCγ in Kidins220^−/− astrocytes (Fig. 1B). We then asked whether ablation of Kidins220 in these cells affects the response to BDNF and TrkB-mediated signaling. At 15 DIV, astrocytes were treated with 50 ng/ml BDNF for 5 or 30 min and levels of phosphorylated forms of MAPK1/2, PLCγ and Akt (P-MAPK, P-PLCγ and P-Akt, respectively) were quantified by western blotting using phospho-specific antibodies (Fig. 1C). Whereas basal levels of phosphorylated MAPK1/2 and Akt were unchanged in Kidins220^−/− astrocytes, we observed a reduction upon BDNF stimulation that reached statistical significance for MAPK1/2 at 30 min. Prolonged treatment with BDNF (up to 2 h) did not further increase levels of P-MAPK and P-Akt in wild-type or Kidins220^−/− astrocytes; moreover, in knock-out (KO) samples, levels of P-MAPK and P-Akt dropped to even lower levels compared with those in wild-type cells exposed to BDNF for 2 h (Fig. S2A). This indicates that absence of Kidins220 affects the magnitude of the BDNF response and not its kinetics. We also detected a significant reduction of the basal PLCγ phosphorylation levels, whereas the percent increase in PLCγ phosphorylation induced by the BDNF was similar in wild-type and Kidins220^−/− astrocytes. A similar trend was observed when BDNF was used at lower concentration (1 ng/ml). However, in this case, the fold increase of MAPK phosphorylation was reduced compared to that when 50 ng/ml BDNF were used for stimulation (Fig. S2B). Of note, we were unable to detect reliable phosphorylation of PLCγ under these experimental conditions (not shown). Whereas astrocytes predominantly express TrkB-T1, there has been some indication that primary astrocytes also express full-length neurotrophin receptors under physiological conditions and/or upon specific stimuli (Climent et al., 2000; Condorelli et al., 1994; Hutton et al., 1992). Although, under our culturing conditions, primary glial cultures are predominantly represented by astrocytes – with only a small percentage of microglial cells (Jaudon et al., 2020), we wanted to confirm that the observed activation of signaling pathways is, indeed, ascribable to astrocytes. Thus, we performed immunostaining experiments using antibodies specifically recognizing TrkB phosphorylated at Tyr516, i.e. the site required to be phosphorylated to activate MAPK, as well as antibodies against GFAP. Indeed, stimulation with BDNF increased fluorescence intensity of pTrkB in GFAP-positive cells, confirming that full-length TrkB activation occurred in astrocytes (Fig. 1D).

Together these results show that, similar to its function in neurons (Arevalo et al., 2004; Cesca et al., 2012), Kidins220 is a crucial regulator of BDNF-dependent TrkB signaling also in astrocytes.

Astrocytes undergo important physiological changes from the embryonic development to the postnatal period, accompanied by variations in the expression pattern of several signaling proteins (Dallerac et al., 2018). To study the developmental changes associated with the BDNF-TrkB system, we prepared primary astrocyte cultures using P0-P1 mouse pups derived by crossing Kidins220^+/lox with Kidins220^+/lox mice, which allowed us to have both wild-type and Kidins220^lox/lox animals in the same litter. We chose this strategy since the full knockout of Kidins220 causes embryonic lethality, thus precluding the possibility of performing postnatal dissection (Cesca et al., 2011, 2012). We first compared the expression levels of full-length and truncated TrkB receptors in wild-type embryonic and postnatal preparations, and found a specific reduction of the full-length isoform in postnatal cultures. This led to a decrease in the ratio of TrkB:TrkB-T1 compared with that in embryonic cells (Fig. 2A). In contrast, expression of p75^NTR was comparable in embryonic and postnatal wild-type cells (Fig. S1B). Wild-type postnatal cultures exposed to 50 ng/ml BDNF showed an appreciable increase of phosphorylated Akt and MAPK1/2, although the fold-increase of P-MAPK1/2 was less than that observed in embryonic cultures (1.2-fold versus 2.5-fold at 30 min; compare Figs 2B and 1C), probably a consequence of reduced TrkB expression. Similar results were observed with a lower BDNF concentration (1 ng/ml; Fig. S3).

Primary Kidins220^lox/lox cultures were infected with lentiviruses expressing either the active form of Cre recombinase to induce depletion of Kidins220 or a functionally inactive form of Cre (ΔCre) as control (Kaeser et al., 2011; see Materials and Methods for details). We refer to these cultures as Kidins220-depleted (Kidins220^lox/lox-Cre) and control (Kidins220^{lox/lox-ΔCre}) cultures. Although Cre-induced depletion of Kidins220 in postnatal preparations yielded reduced levels of full-length and truncated TrkB, the TrkB to TrkB-T1 ratio remained unaltered (Fig. 3A). Levels of p75^NTR were also unchanged in the absence of Kidins220 (Fig. S1C). Both results are in full agreement with those obtained on embryonic cultures. We then analyzed the signaling capability of postnatal cultures in the presence or absence of Kidins220. The basal expression levels of MAPK1/2, PLCγ and Akt were not affected by depletion of Kidins220 (Fig. 3B). Moreover, in contrast to the data obtained in embryonic cultures, removal of Kidins220 in postnatal cultures did not affect activation of BDNF-induced phosphorylation of MAPK1/2, PLCγ and Akt (Fig. 3C and Fig. S4).

Altogether, the data presented in Figs 2 and 3 show (i) reduced protein levels of full-length TrkB and reduced activation of kinase pathways upon exposure to BDNF in wild-type postnatal astrocytes compared to embryonic cells; and (ii) a further reduction of both TrkB and TrkB-T1 levels in the absence of Kidins220. However the limited activation of kinase pathways induced by BDNF was not impacted upon depletion of Kidins220.

We subsequently investigated the role played by Kidins220 in BDNF-induced [Ca²⁺]_i signaling by monitoring BDNF-induced [Ca²⁺]_i transients in astrocytes (Rose et al., 2003). In embryonic cultures, very few cells responded to BDNF stimulation – 9.8% for wild-type and 7.4% for Kidins220^−/− (Fig. 4A) and responsive cells displayed modest [Ca²⁺]_i variations (Fig. 4B,C). In contrast, ∼30–35% of cells in postnatal cultures responded with sizable [Ca²⁺]_i peaks to BDNF (Fig. 4D–F). Whereas the percentage of responding cells was not affected by Kidins220 depletion (Fig. 4D), the average amplitude of BDNF-induced [Ca²⁺]_i transients was significantly reduced in absence of Kidins220 (Fig. 4E and F; ^#P=0.0432 genotype effect). To investigate the contribution of the different isoforms of TrkB in this process, we repeated the same experiments in the presence of either the tyrosine kinase blocker K-252a (Cesca et al., 2012) to specifically inhibit the full-length TrkB isoform, or the TrkB antagonist ANA-12 (Cazorla et al., 2011) to block both TrkB and TrkB-T1. Pre-incubation of cells with K-252a did not affect the percentage of BDNF-responding cells (Fig. 4D, bars on shaded background). However, the amplitude of the observed [Ca²⁺]_i influx was reduced independently of the presence of Kidins220 (Fig. 4F and G; genotype effect: ^#P=0.0432; K-252a treatment effect: ***P<0.0001; treatment×genotype interaction: P=0.2197, see legend for more details), suggesting that the BDNF-induced [Ca²⁺]_i variations involved full-length TrkB receptors. BDNF-stimulation in the presence of ANA-12 did not induce any response, confirming that the residual signal was attributable to TrkB-T1 activation (Fig. 4H). Furthermore, the observed response was completely dependent on PLCγ activity, as we did not observe BDNF-induced [Ca²⁺]_i transients in the presence of the specific PLCγ inhibitor U73122 (Fig. 4I).

Altogether, these data show that BDNF induced modest [Ca²⁺]_i events in embryonic astrocytes, whereas larger [Ca²⁺]_i responses are elicited in postnatal cells. Postnatal [Ca²⁺]_i transients are fully dependent on TrkB receptors, as shown by complete inhibition of TrkB and TrkB-T1 in the presence of ANA-12 (Fig. 4H). Interestingly, part of this response is mediated by the full-length receptor, as shown by the partial reduction of amplitude in the presence of K-252a (Fig. 4F, K-252a treatment effect). Moreover, our results show that Kidins220 plays an active role in this process, as in its absence the amplitude of [Ca²⁺]_i events is reduced (Fig. 4F, genotype effect).

To shed light on the functional effects of BDNF signaling in astroglia, we analyzed the transcriptional profile of cultures treated with 1 ng/ml BDNF for three days. We focused on four genes whose expression is crucial for astrocytes physiology, i.e. the excitatory amino acid transporter 2 (Slc1a2, hereafter referred to as GLT-1), the ATP-sensitive inward rectifier potassium channel 10 (Kcnj10, hereafter referred to as Kir4.1), aquaporin 4 (Aqp4) and gap junction alpha-1 protein (Gja1, hereafter referred to as Cnx43) (Fig. 5A). Whereas the expression of the genes encoding GLT-1, Kir4 or Cnx43 was insensitive to BDNF, at least within the time frame of our experiment, transcription of Aqp4 was selectively reduced upon prolonged stimulation (72 h) with BDNF (*P=0.0153, treatment effect). Although Kidins220 deficiency did not alter the transcription response to BDNF, it selectively and markedly reduced the protein levels of Kir4.1 independently of BDNF (^##P=0.0067, genotype effect), which was confirmed by western blot analysis (Fig. 5B).

These data suggest that (i) prolonged exposure to BDNF alters the expression of specific astrocyte genes, and that (ii) Kidins220 plays a role in the maintenance of astrocyte homeostasis and K⁺ buffering function by modulating the expression of Kir4.1.

One of the main functions of astrocytes is to provide neurons with metabolic support, both under physiological conditions and in situations of high energetic demand. Amongst the numerous astrocyte-derived molecules involved in this process, lactate plays a prominent role both as a metabolic and as a signaling molecule (Magistretti and Allaman, 2015). To assess the levels of glycolytic versus oxidative phosphorylation metabolism, we quantified the amount of lactate in the medium of embryonic and postnatal astrocytes, in the presence and absence of Kidins220 (Fig. 6A). Interestingly, we found that, the production of lactate in wild-type postnatal cells is increased compared with that in embryonic cells, suggesting a predominant glycolytic mechanism. Remarkably, Kidins220^−/− embryonic cells produce three times more lactate compared to wild-type astrocytes, suggesting impaired mitochondrial metabolism, whereas at the postnatal stage lactate levels in Kidins220-depleted cells are comparable to those in wild-type cultures.

Altogether, these data show that the astrocytic metabolism changes during development by way of increasing the glycolytic power, and reveal a novel and previously unknown role of Kidins220 in cell energy metabolism.

The main findings of this work can be summarized as follows. (i) Cultured primary astrocytes predominantly express the TrkB-T1 truncated isoform of the receptor but also express small amounts of full-length signaling-competent TrkB, as shown through BDNF-induced activation of the MAPK pathway; however, expression of TrkB declines in postnatal cells. (ii) Astrocyte cultures show a stage-dependent shift in their response to BDNF, i.e. embryonic cells respond mostly by activation of intracellular signaling pathways (MAPK1/2, PCLγ and Akt) and are unable to induce significant BDNF-dependent [Ca²⁺]_i transients. Postnatal astrocytes in culture, however, respond to BDNF with the induction of robust [Ca²⁺]_i transients and the limited activation of intracellular signaling cascades. (iii) The scaffold protein Kidins220 participates in TrkB signaling within both embryonic and postnatal astrocytes, and modulates the transcription of fundamental astrocyte genes, such as that encoding the inward rectifier K⁺ channel Kir4.1. A schematic representation of our results is shown in Fig. 6B.

Extensive literature supports the notion that TrkB-T1 is the main isoform expressed in astroglia (Fenner, 2012; Holt et al., 2019); however, experimental evidence also points at a limited expression of the full-length receptor, at least in vitro (Climent et al., 2000; Condorelli et al., 1995). Moreover, reactive astrocytes express the full-length TrkB receptor upon CNS injury (McKeon et al., 1997) and in the case of chronic diseases (Soontornniyomkij et al., 1998; Stadelmann et al., 2002), raising the question of whether the astrocyte expression of full-length TrkB under pathological conditions recapitulates intracellular pathways that are normally active during development. Re-activation of developmental programs is, indeed, a known feature of reactive astrocytes (Buffo et al., 2008) and of astrocyte-derived tumors, such as gliomas (Galli et al., 2004; Ignatova et al., 2002; Silver and Steindler, 2009), and the BDNF-TrkB pathway is over-activated in several brain tumors (Meng et al., 2019). Consistent with previous observations (Climent et al., 2000), our data show that the expression of full-length TrkB decreases from embryonic to postnatal astrocytes, supporting the idea that the intracellular pathways activated by this receptor could be active during embryogenesis and are re-activated in the presence of neural pathologies. This finding is potentially relevant in that it shows that cultured astrocytes do retain – to some extent – the identity of the tissue from which they were initially isolated (Schwartz and Nishiyama, 1994; Wu et al., 1998).

Regarding the signaling pathways activated through BDNF in astrocytes, our data indicate activation of the canonical MAPK1/2, PCLγ and Akt pathways is predominant in embryonic cultures. The same cascades are activated to a reduced extent in postnatal cells, which is in agreement with the reduced expression of full-length TrkB. Our embryonic and postnatal cultures are predominantly composed of astrocytes, with only a small (<10%) proportion of other glial cells, mostly microglia (Jaudon et al., 2020). Although we cannot exclude a small proportion of activated pathway proteins observed in western blot experiments being derived from cells other than astrocytes, the high levels of phosphorylated pathway proteins and the increased amounts of phosphorylated TrkB in GFAP-positive cells in response to BDNF treatment – revealed by western blotting and fluorescence microscopy, respectively – indicate that the observed changes in intracellular signaling pathways are predominantly attributable to astrocytes. The mechanisms that underlie BDNF-induced [Ca²⁺]_i transients become fully functional in postnatal astrocytes; indeed, only modest BDNF-induced [Ca²⁺]_i elevations are observed in embryonic cultures. We hypothesize that such a limited response is ascribable to a still non-functional coupling between Ca²⁺ receptors and channels, on the endoplasmic reticulum and on the plasma membrane, with these together coordinating the mechanism of store-operated Ca²⁺ entry (SOCE) in mature astrocytes (Okubo, 2020). The [Ca²⁺]_i transients observed upon BDNF administration in postnatal cells comprise two components. The first, which accounts for ∼40% of the observed signal, is mediated by full-length TrkB and sensitive to pre-incubation with K-252a. Complete ablation of [Ca²⁺]_i transients observed in the presence of the pan-TrkB inhibitor ANA-12 reveals that the second, and predominant, component depends on TrkB-T1, as previously reported (Rose et al., 2003). The complete dependence of [Ca²⁺]_i variations on PLCγ suggests that pathways activated by both receptors converge upon SOCE. Whereas only 40% of postnatal astrocytes showed responsiveness to BDNF in terms of [Ca²⁺]_i transients, we found phosphorylated full-length TrkB in the majority of GFAP-positive cells within embryonic cultures. Our Ca²⁺-imaging data confirmed that the capability of astrocytes to initiate BDNF-induced [Ca²⁺]_i transients varies, depending on the degree of cell differentiation. Thus, we propose that expression of full-length TrkB is a general feature of embryonic cells. Whether BDNF-responsive astrocytes in postnatal cultures represent a specific subpopulation endowed with distinct physiological functions is a relevant issue that needs to be addressed in future investigations. Interestingly, prolonged treatment with BDNF reduced gene expression of Aqp4, indicating that the neurotrophin is capable to induce long-term changes in astrocytes, potentially impacting on their physiology and on their capability to maintain nervous system homeostasis. In contrast to the changes observed regarding the levels of TrkB, levels of p75^NTR remained constant in embryonic and postnatal cells. Thus, changes in the amount of p75^NTR seem more likely to be linked to injury responses (Cragnolini et al., 2018) than to the developmental stage of the cell.

The scaffold protein Kidins220 is highly expressed in the nervous system during embryonic development and plays a fundamental role in modulating neuronal survival and maturation. Indeed, its complete ablation leads to embryonic lethality with widespread apoptosis in the central and peripheral nervous system (Cesca et al., 2011, 2012). The function of Kidins220 has been extensively studied in neurons, where it contributes to several fundamental processes, such as neuronal differentiation (Higuero et al., 2010), excitability (Cesca et al., 2015) and synaptic plasticity (Scholz-Starke et al., 2012), mostly by mediating BDNF stimuli through its interaction with TrkB and p75^NTR (Arevalo et al., 2006, 2004; Kong et al., 2001). In the absence of Kidins220, expression of both TrkB isoforms is reduced in embryonic as well as in postnatal astrocyte cultures. In embryonic cultures, Kidins220^−/− astrocytes are characterized by reduced sustained activation of the MAPK pathway (Arevalo et al., 2004; Cesca et al., 2012). This suggests that – similar to modulation in neurons – Kidins220 modulates MAPK activation downstream of full-length TrkB in astrocytes. In postnatal astrocytes, Kidins220 affects both TrkB- and TrkB-T1-induced [Ca²⁺]_i signaling, as both isoforms are reduced in Kidins220-depleted cells. It should be noticed that the interaction between Kidins220 and TrkB involves the transmembrane domain of the receptor (Arevalo et al., 2004), which is shared between the two receptor isoforms. Our gene expression analysis revealed reduced expression of Kir4.1 in Kidins220-depleted astrocytes. In astrocytes, Kir4.1 channels exist as homo-tetramers comprising Kir4.1 subunits as well as hetero-tetramers comprising Kir4.1 and Kir5.1 subunits (Hibino et al., 2004; Patterson et al., 2021). In our quantitative real-time PCR (RT-qPCR) and western blot analysis, we quantified monomeric Kir4.1, which is likely to reflect the abundance of both homo- and hetero-tetramers. Interestingly, similar results have been recently described in cultured astrocytes lacking TrkB-T1 (Aroeira et al., 2015; Holt et al., 2019; Ohira et al., 2007; Vaz et al., 2011). This further supports the notion that Kidins220 and TrkB receptors are part of the same signaling system. Together with the increased expression of transient receptor potential cation channel subfamily V member 4 (TRPV4) previously described by us (Jaudon et al., 2020), this observation suggests that the Kidins220 deficiency subjects astrocytes to a general stress reaction and aids to explain the developmental and activity impairments observed for wild-type neurons when co-cultured with Kidins220^−/− cells (Jaudon et al., 2020).

Our data on lactate production raise a number of interesting considerations. For what concerns wild-type cells, our findings are in line with evidence that, in mature astrocytes, production of lactate through anaerobic glycolysis plays a predominant role in astrocyte metabolism (reviewed by Magistretti and Allaman, 2015). For what concerns Kidins220 KO cells, the present findings together with our previous data of reduced ATP secretion in embryonic Kidins220 KO astrocytes (Jaudon et al., 2020), suggest that in KO embryonic astrocytes the balance between glycolysis and oxidative phosphorylation is towards lactate production. This could be indicative of mitochondrial dysfunctions. The mechanisms underlying this effect are presently unknown, and the role of Kidins220 in mitochondria physio-pathology – and, generally, in cellular metabolism – is completely unexplored. The induction and modulation of [Ca²⁺]_i transients could be a major player, especially given the strict crosstalk between endoplasmic reticulum and mitochondria regarding [Ca²⁺]_i homeostasis (Verkhratsky et al., 2018).

Overall, our work contributes to understanding the complex and debated mechanisms of neurotrophin signaling in astrocytes, showing that – at least in vitro – the full-length and the truncated TrkB receptor are both competent to activate distinct intracellular signaling pathways at distinct stages of astrocyte development. Moreover, we identified Kidins220 as one of the cellular components that endows TrkB and TrkB-T1 signaling specificity. In future studies we will address how these findings translate in vivo, under physiological and pathological conditions. In recent years, mutations in the KIDINS220 gene have been associated with severe neurodevelopmental pathologies, whose main symptoms are intellectual disability and spastic paraplegia (Cesca et al., 2018; El-Dessouky et al., 2020; Jacquemin et al., 2020; Josifova et al., 2016; Mero et al., 2017; Yang et al., 2018; Zhao et al., 2019). Within this context, the observation that Kidins220 controls astrocyte metabolism during embryonic development is instrumental to tackle the pathogenic mechanisms that underlie Kidins220-dependent neurodevelopmental diseases (i.e. SINO syndrome and associated pathologies). A better understanding of the physiological functions of neurotrophins and Kidins220 in neurons and glial cells is instrumental to identify the molecular pathways that lead to this severe neurological disorder.

All embryos used in this study were obtained from crosses of Kidins220^+/− mice (Cesca et al., 2012; Scholz-Starke et al., 2012) in the C57BL/6 background. Mice were mated overnight and separated in the morning. The development of the embryos was timed from the detection of a vaginal plug, which was considered day 0.5. Postnatal cultures were prepared from P0-P1 pups obtained by crossing Kidins220^+/lox mice on the C57BL/6 background. All experiments were carried out according to the guidelines established by the European Community Council (Directive 2010/63/EU of 22 September 2010) and were approved by the Italian Ministry of Health.

The following primary antibodies were used: rabbit polyclonal anti-Kidins220 (1:500; GSC16, #AB34790, Abcam), rabbit polyclonal anti-TrkB (1:1500; #07-225, Millipore), rabbit polyclonal anti-p75^NTR (1:2000; #G323A, Promega), rabbit monoclonal anti-phosphorylated TrkB (1:1000; Tyr516, #4619, Cell Signaling), rabbit monoclonal anti-GAPDH (1:10,000; 14C10, #2118, Cell Signaling), rabbit monoclonal anti-phosphorylated MAPK1/2 (1:2000; Thr202/Tyr204, #4377, Cell Signaling), rabbit polyclonal anti-MAPK1/2 (1:2000; #06-182, Millipore), rabbit monoclonal anti-phosphorylated Akt1, Akt2 and Akt3 (1:1000; Ser473, #4058, Cell Signaling), rabbit polyclonal against Akt1, Akt2 and Akt3 (1:1000; #9272, Cell Signaling), rabbit polyclonal anti-phosphorylated PLCγ1 (1:500; Tyr783, #2821, Cell Signaling), rabbit polyclonal anti-PLCγ1 (1:1000; #2822, Cell Signaling), mouse monoclonal anti-glial fibrillary acidic protein (1:500; GFAP, #G3893, Sigma-Aldrich), anti Kir4.1 (1:3000; #APC-035 Alomone Labs) and anti-calnexin (1:5000; #ADI-SPA-860 Enzo Life Sciences).

Secondary antibodies for western blot analysis were ECL Plex goat anti-rabbit IgG-Cy5 (1:2500; PA45012, GE Healthcare), ECL Plex goat anti-mouse IgG-Cy3 (1:2500; PA43009, GE Healthcare), HRP-conjugated goat anti-rabbit antibodies (1:5000; #31460, ThermoFisher Scientific) and HRP-conjugated goat anti-mouse antibodies (1:5000; #31430, Thermo Fisher Scientific). Fluorescent secondary antibodies for immunocytochemistry were from Molecular Probes (Thermo Fisher Scientific; Alexa Fluor 488, 1:500, #A11029; Alexa Fluor 647, 1:500, #A21450). Hoechst 33342 (1 mg/ml; #B2261, Sigma-Aldrich) was used to stain nuclei.

E18.5 or P0-P1 cortices were dissected in ice-cold PBS, incubated with trypsin 0.25% and 1 mg/ml DNase I for 30 min at 37°C, and mechanically dissociated. Cells were then re-suspended and plated on poly-D-lysine-coated flasks or glass coverslips, in Minimum Essential Medium (MEM) containing 10% FBS, 2 mM glutamine, 33 mM glucose and antibiotics (astrocyte culture medium). After 24 h, the medium was removed and replaced with fresh culture medium. After 1 week, half of the medium was replaced with fresh culture medium. Embryonic and postnatal cultures comprised ∼90% astrocytes and were totally devoid of neurons (Jaudon et al., 2020).

HEK293T (ATCC CRL-3216) cells were maintained in Iscove's Modified Dulbecco's Medium (IMDM) supplemented with 10% FBS, 2 mM glutamine, 100 U/ml penicillin and 0.1 mg/ml streptomycin in a 5% CO₂ humidified incubator at 37°C. Cells were routinely tested and resulted mycoplasma free. Cells were transfected with the Δ8.9 encapsidation plasmid, the VSVG envelope plasmid and the pLenti-PGK-Cre-EGFP or pLenti-PGK-ΔCre-EGFP plasmids (Kaeser et al., 2011) using the calcium phosphate method. Transfection medium was replaced with fresh medium after 16 h. Supernatants were collected between 36 and 48 h after transfection, centrifuged to remove cell debris, passed through a 0.45 µm filter and ultracentrifuged 2 h at 20,000 g at 4°C. Viral pellets were re-suspended in PBS, aliquoted and stored at −80°C until use.

Confluent postnatal cultures were trypsinized and seeded on 6-well plates or glass coverslips coated with poly-D-lysine for subsequent experiments. Cells were infected 3 days later with lentiviruses encoding catalytically dead (ΔCre) or active Cre recombinase with the lowest infectious dose capable of transducing ≥95% of cells (dilution range 1:500 to 1:800) and used for experiments ≥7 days after transduction.

Cells were washed once in ice-cold PBS and lysed in RIPA buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 2 mM EDTA, 1% NP40, 0.1% SDS) plus protease and phosphatase inhibitors (complete EDTA-free protease inhibitors, Roche Diagnostic; serine/threonine phosphatase inhibitor and tyrosine phosphatase inhibitor, Sigma). After centrifugation at 16,000 g for 15 min at 4°C, protein concentration was quantified using the BCA Protein Assay kit (ThermoFisher Scientific). SDS-PAGE and western blotting were performed using precast 4–12% NuPAGE Novex Bis-Tris Gels (Invitrogen). After incubation with primary antibodies, membranes were incubated with fluorescent secondary antibodies and revealed by a Typhoon Variable Mode Imager (GE Healthcare) or with HRP-conjugated secondary antibodies and ECL Prime Western Blotting System (#RPN2106, GE Healthcare) and imaged using a ChemiDoc imaging system (Biorad). Immunoreactive bands were quantified using ImageJ software. The fold-change of activation was calculated by first calculating the intensity ratio of phosphorylated bands to the total amount of corresponding protein for each sample, and then by dividing the ratios of phosphorylated to total treated samples by the ratio of phosphorylated to total untreated control samples.

Cells were fixed in 4% PFA for 15 min, then washed in PBS. Cells were subsequently permeabilized with 0.2% Triton X-100 in PBS for 10 min at room temperature (RT) then incubated with primary antibodies diluted in PBS 1% BSA overnight at 4°C or 2 h at RT. After five washes in PBS, cells were incubated with fluorescent secondary antibodies diluted in PBS 1% BSA. After washes, coverslips were mounted with Mowiol.

Confocal stacks were acquired with a Leica SP8 using a 40× oil immersion objective (NA 1.40), with 1 µm between optical sections. Images were analyzed using ImageJ software. The maximal fluorescence intensities of in-focus stacks were z-projected and the resulting images were automatically thresholded. Regions of interests were manually drawn around the borders of GFAP-positive cells in the red channel; fluorescence intensity was measured in the green channel (P-TrkB). Quantification was performed on 18–20 fields per condition from three independent cell preparations. Data are reported as the average P-TrkB fluorescence intensity of all GFAP-positive cells in each field.

Three days after lentiviral infection, Kidins220^lox/lox astrocytes cultures were serum starved overnight then treated with 1 ng/ml BDNF on the following morning or left untreated. Treatment was repeated on the next 2 days and total RNA was extracted on the third day with the QIAzol lysis reagent (Qiagen). The corresponding cDNAs were prepared by reverse transcription of 1 µg of RNA using the SuperScript III First-Strand Synthesis System (Invitrogen) with an oligo-dT primer according to the manufacturer's instructions. The resulting cDNAs were used as a template for RT-qPCR using a CFX96 Real-Time PCR Detection System (Biorad) with a SYBR Green master mix (Qiagen). Thermal cycling parameters were 5 min at 95°C, followed by 40 cycles of 95°C for 15 s, and 60°C for 45 s. The relative quantification in gene expression was determined using the ΔΔCt method. Data were normalized to Transferrin receptor protein 1 (TRFR), TATA-box-binding protein (TBP), and Tubulin beta-2A (TUBB2) by the multiple internal control gene method with GeNorm algorithm (Vandesompele et al., 2002). Sequences of the primers used are listed in Table S1.

Lactate contents were measured using an L-Lactate assay kit (#MAK064, Sigma) according to the manufacturer's instructions. Culture medium was collected from confluent cultures, filtered through a 30 kDa cut-off spin filter and stored at −80°C until processing. Reaction reagents were prepared as instructed and mixed with 50 µL of the samples, incubated at room temperature for 30 min protected from light, and measured at OD 570 nm. Lactate concentration was normalized to the number of cells.

Data are presented as the means±standard error of the means (±s.e.m.) throughout the text. Data distribution was assessed using the D'Agostino-Pearson omnibus normality test. When comparing two groups, unpaired two-sided Student's t-test and one-sample t-test were used; equality of variances were assessed by using the F test. When more than two groups were compared, two-way ANOVA followed by Tukey's post-hoc multiple comparison test was performed to assess significance as indicated in figure legends; equality of variances was assessed by using the Brown-Forsythe's and Bartlett's test. Alpha levels for all tests were 0.05% (95% confidence intervals). No statistical methods were used to predetermine sample sizes; however, sample sizes used for the experiments described here (indicated in figure legends) were similar to those previously reported in the literature for similar experiments. The ROUT method with Q=1% was used to identify outliers for exclusion from analysis. All statistical procedures were performed using GraphPad Prism 7 software (GraphPad Software, Inc).

We kindly acknowledge Drs A. Mehilli, D. Moruzzo and R. Navone for help with maintenance, breeding and genotyping of the Kidins220 mouse strains; and Drs R. Ciancio and I. Dallorto for the administrative support.

The peer review history is available online at https://journals.biologists.com/jcs/article-lookup/doi/10.1242/jcs.258419