Activity of Human-Specific Interlaminar Astrocytes in a Chimeric Mouse Model of Fragile X Syndrome

Abstract

Astrocytes, a subtype of glial cells, have multiple roles in regulating neuronal development and homeostasis. In addition to the typical mammalian astrocytes, in the primate cortex, interlaminar astrocytes are located in the superficial layer and project long processes traversing multiple layers of the cerebral cortex. Previously, we described a human stem cell based chimeric mouse model where interlaminar astrocytes develop. Here, we utilized this model to study the calcium signaling properties of interlaminar astrocytes. To determine how interlaminar astrocytes could contribute to neurodevelopmental disorders, we generated a chimeric mouse model for Fragile X syndrome (FXS). We report that FXS interlaminar astrocytes exhibit hyperexcitable calcium signaling and are associated with dendritic spines with increased turnover rate.

1. Introduction

Astrocytes, a subtype of glial cells in the central nervous system (CNS), are crucial for maintaining CNS homeostasis [1]. They are intricately involved in neural circuits and influence various aspects of CNS function, including the regulation of neuronal development, synaptic transmission, ionic homeostasis, neurotransmitter recycling, metabolic regulation, and the maintenance of the blood–brain barrier [2,3,4]. Astrocyte excitability is manifested as elevations of cytosolic calcium (Ca²⁺). Astrocyte Ca²⁺ elevations can occur spontaneously as intrinsic oscillations in the absence of neuronal activity, can be triggered by neurotransmitters released during synaptic activity, and can occur in response to sensory stimulation [5]. Increasing evidence highlights the heterogeneity of astrocyte functions, which arise from diverse morphologies and transcriptomic profiles across different brain regions [6]. However, much of this knowledge comes from studies on rodent astrocytes, which differ from their human counterparts in several key aspects [7,8,9], leaving the structural and functional properties of human astrocytes largely unexplored.

In comparison to rodents, human astrocytes exhibit distinct genomic profiles, larger territories, more complex morphologies, faster intracellular Ca²⁺ signaling, and morphological features unique to primates [8,9,10]. Astrocyte subtypes are regionally distributed within the brain, including protoplasmic astrocytes in the gray matter, fibrous astrocytes in the white matter, and other types such as velate astrocytes, perivascular astrocytes, Müller cells, and Bergmann glia [1]. Additionally, higher-order primates and humans possess astrocyte subtypes like interlaminar and varicose projection astrocytes, which are not found in rodents [1,8,11,12,13]. In the human cortex, interlaminar astrocytes (ILAs) reside in layer I and extend long processes that traverse multiple cortical layers, terminating in layers III/IV [12,14]. While rudimentary interlaminar astrocytes can also be found in the rodent cortex, their processes are shorter and do not extend beyond layer I [12]. The functional roles of primate-specific interlaminar astrocytes in the development, maintenance, and function of cortical neural circuits remain poorly understood.

We recently developed a chimeric mouse model with human-induced pluripotent stem cell (hiPSC)-derived astrocytes, in which ILAs are present in the mouse cerebral cortex [15]. This model offers the first opportunity to investigate the functional properties of ILAs. In this study, we conducted Ca²⁺ imaging of ILA processes both in brain slices and in vivo, assessing their responses to canonical neurotransmitters. To assess a potential contribution of ILAs to neurodevelopmental disorders, we also engrafted astrocytes derived from hiPSCs of individuals with Fragile X syndrome (FXS), the most common inherited form of autism spectrum disorder. Given our previous findings that cultured astrocytes derived from FXS hiPSCs exhibit hyperexcitable Ca²⁺ signaling [16], we compared the Ca²⁺ signaling between FXS and control ILAs. Additionally, using multiphoton imaging of dendritic spines, we observed that dendrites near FXS ILAs have higher spine turnover rates compared to those near control ILAs, suggesting the FXS ILAs contribute to altered synaptic plasticity in FXS.

2. Results

2.1. Generation of hiPSCs-Astrocyte Chimeric Mice

We differentiated human-induced pluripotent stem cells (hiPSCs) to neural progenitor cells (NPCs), and subsequently astrocytes [15,17]. To visualize the hiPSC-astrocytes (hi-Astrocytes) in the chimeric mouse brain and to measure astrocyte Ca²⁺ signaling, we generated RFP or mScarlet and GCaMP6f-expressing astrocytes. We previously demonstrated that prior to engraftment, RFP-expressing hi-Astrocytes showed robust expression of canonical astrocyte markers [15]. Two-site injection resulted in the widespread distribution of hi-Astrocytes in the frontal cortex (Figure 1A). At what we presume to be sites of injection, both ILA and deeper protoplasmic astrocytes are observed, and farther away, cells are mainly confined to layer I and likely comprise both pial and subpial ILAs (Figure 1B).

2.2. Calcium Signaling Properties of ILAs

The dynamic Ca²⁺ properties of ILAs and their long processes have not been previously examined. We therefore imaged slices at 6 months post-engraftment in cells that expressed GCaMP6f and the structural marker mScarlet. At this age Ca²⁺ signals in both cell bodies at the pial surface and the long ILA processes that project from the pial surface could be examined (Figure 2A,B). To capture as much of the long ILA processes as possible, we performed volumetric 2-photon imaging of a 20-micron volume at 1 Hz and imaged both spontaneous and agonist-induced changes (Figure 2). We observed very little spontaneous activity with none of the ILA somas exhibiting activity in the absence of an agonist. We next asked if hiPSC-derived ILA cells are responsive to the canonical transmitters that have been examined extensively in mice; ATP (100 µM) to activate purinergic receptors, norepinephrine (NE, 50 µM) to activate noradrenergic receptors, and carbachol (CA, 50 µM) to activate cholinergic receptors. We found that 56% and 44% of processes responded to ATP (N = 141) and NE (N = 100), respectively (Chi-Square, p = 0.043). Interestingly, none of the ILAs responded to carbachol. These data suggest that engrafted hi-Astrocytes that develop into ILAs express purinergic and adrenergic receptors and the molecular machinery for Ca²⁺ signaling. However, the lack of response to carbachol suggests low or absent expression of cholinergic receptors on ILAs.

We next characterized the properties of the Ca²⁺ signals in the somas and the processes. ILA somas, from engrafted control hi-Astrocytes, had robust Ca²⁺ responses to both ATP and NE that lasted for tens of seconds (Figure 3A, Supplemental Data Movie S1). We did not observe a difference in the peak amplitude (dF/F), area under the curve or the duration of the ATP- and NE-induced responses in control ILA somas (ATP N = 36 and NE N = 27 somas, multiple Mann–Whitney tests p > 0.05). Since ILAs do not appear to have fine processes that are found on protoplasmic astrocytes, we only considered event sizes larger than 3 μm² and found that following ATP or NE application, Ca²⁺ events spanned a range of 3–105 μm²; however, the distribution of ATP-induced Ca²⁺ events was shifted towards larger events compared to NE-induced events (Figure 3B, Kolmogorov–Smirnov test, p = 0.025). The maximum event size is constrained by the length of the outlined processes that are observed within the slices. However, the larger event areas with ATP were not due to longer regions of interest, as in fact they were longer in the NE treated slices (ATP: 91.99 ± 4.28 µm (N = 28); NE: 114 ± 5.3 µm (N = 31), unpaired t-test p = 0.002). We next characterized the Ca²⁺ signaling properties based on event sizes dividing them into small (3–10 μm²) and large events (>10 μm²) (Figure 3C). Despite the difference in the event areas, the dynamic properties of the ATP- and NE-induced Ca²⁺ events in control cells were not different in any of the parameters examined (Figure 3C). These studies demonstrate that the long ILA processes exhibit Ca²⁺ signaling in response to canonical agonists.

To determine if the Ca²⁺ signaling observed in ILA processes occur in the intact brain, we performed 2-photon in vivo imaging of hi-Astrocyte engrafted mice through a cranial window in 6-month old awake head-restrained conditions. Similar to the observed Ca²⁺ activity in slices, we observed that long, thin processes in cortical Layer I exhibited robust Ca²⁺ responses in vivo (Figure 2C). The activity is likely to have been induced by the movement of the mouse (Supplemental Data Movie S2).

2.3. ILA Process Length Is Not Significantly Altered in FXS

Impairments in astrocyte function are increasingly associated with neurodevelopmental disorder [18]. However, how ILAs that are specific to humans and non-human primates are altered in neurodevelopmental disorders has not been examined before. We have utilized the human astrocyte mouse chimera model with astrocytes derived from FXS hiPSCs or human embryonic stem cells (hESCs) with deletion of FMR1 to determine if ILA development and function is altered. We have previously described the gradual development of ILA processes in the engrafted mice with few processes observed at 3 months and full “palisade-like” processes observed at 9 months of age [15]. Here we first asked if process length was altered in FXS ILAs. Due to the large number of processes in the “palisade-like” distribution, we were unable to reliably track the processes of individual cells. Instead, we measured the distance of the “interlaminar palisade” from the pial surface in FX11-9u (CTR) and FX11-7 (FXS) hi-Astrocyte engrafted mice at 3 and 9 months of age (Figure 4B). While the processes displayed a significant age factor (F (2, 40) = 16.55, p < 0.0001) and FXS ILA processes tended to be longer, the genotype factor was not significant (F (1, 40) = 3.129, p = 0.08) nor was there a significant interaction (F (2, 40) = 0.7415, p = 0.48) (3 months: 145.00 ± 38.81 and 228.9 ± 29.90 μm, 9 months: 572.7 ± 81.49 and 673.4 ± 89.44 μm). Thus, we conclude that there was no gross alteration in the development of ILAs in FXS.

2.4. FXS ILAs Have Increased Calcium Signaling

We next compared the dynamic Ca²⁺ properties of engrafted astrocytes derived from control (3 lines, same data as in Figure 3) and FXS (2 lines) human stem cells. At 6 months of age, we observed that the somas of FXS ILAs had an approximately 60% and 100% increase in the duration of the response to ATP and NE, respectively (Figure 5A and Figure 6A, ATP, N = 36 control and N = 33 FXS somas, p = 0.0002; NE, N = 27 control and N = 12 FXS somas, p = 0.008. Multiple Mann–Whitney tests). We did not observe a difference between the peak amplitude (dF/F) of the soma responses to either ATP or NE between control and FXS cells (Figure 5A and Figure 6A). Interestingly, analysis of hi-Astrocytes from a subset of lines after 4 months of engraftment revealed multi-peak Ca²⁺ transients in most cells in response to bath-application of 100 µM ATP (Figure S1A). Unlike the increased duration observed at 6 months, at 4 months we observed increased peak amplitude in engrafted FXS hi-Astrocytes (Figure S1B, CTR: 0.77 ± 0.087 dF/F, FXS: 1.18 ±.065 dF/F, p = 0.0023, N = 6–8 slices from 4–5 mice in each group). These data demonstrate that FXS derived astrocytes that develop in vivo, exhibit hyperexcitability of Ca²⁺ in response to multiple agonists.

We next examined if the Ca²⁺ hyperexcitability observed in FXS somas is also observed in processes. Analysis of the ILA processes revealed no difference in the size of the area of ATP- and NE-induced Ca²⁺ events detected by AQuA between CTR and FXS (Figure 5B and Figure 6B). However, analysis of small and large ILA events determined that while most parameters were similar between CTR and FXS (Figure 5C and Figure 6C), a 62% increase in the duration of large ILA events in response to ATP was observed in FXS (CTR: N = 19, FXS = 26 processes, p < 0.05). Overall, these data indicate hyperactive calcium signaling in FXS ILAs.

2.5. Increased Dendritic Spine Turnover in Mice Engrafted with FXS ILAs

Cumulative evidence showed that astrocytes regulate spine formation and elimination to maintain dynamic and precise neuronal connections. Previous studies identified increased spine dynamics in the Fmr1 KO mice [19,20,21]. Several studies described the contribution of astrocytes to the neuronal abnormalities in the mouse model of FXS [22,23,24,25]. Thus, we asked whether human astrocytes contribute to the altered spine plasticity in FXS by repeatedly imaging the same neuronal dendritic segments in vivo in the chimeric mice. Excitatory neurons in the chimeric mice were labeled with eGFP through viral injection [26]. Dendritic spines in proximity of RFP-expressing hi-Astrocytes (distance < 20 µm) were imaged through the cranial window twice with 4-day intervals (Figure 7A). Dendritic spine density on dendrites in the vicinity of FXS hi-Astrocytes showed a trend toward increased density (Figure 7B, CTR: 2.54± 0.21, FXS: 3.48 ± 0.42, p = 0.082). Though there was only a trend toward an increased percentage of newly formed dendritic spines in the FXS group compared with that in the CTR group (Figure 7C, CTR: 20.89 ± 5.78%, FXS: 34.86 ± 6.34%, p = 0.16, 16–28 regions from N = 5 mice in each group), the percentage of spine elimination on dendrites in the FXS group was higher than that in the CTR group (Figure 7D, CTR: 19.11 ± 199%, FXS: 29.1 ± 1.33%, p = 0.003), as was the turnover rate (TOR) [20,27] (Figure 7E, CTR: 0.21 ± 0.04, FXS: 0.32 ± 0.03, p = 0.037). These results demonstrate an increase in host dendritic spine plasticity attributed to engrafted FXS human astrocytes, indicating the role of astrocytes in the spine dynamics in FXS.

3. Discussion

We used hi-Astrocyte chimeric mice to characterize the Ca²⁺ signaling properties of ILAs and show Ca²⁺ increases in both the cell bodies and processes. We also for the first time report the capacity of engrafted ILA processes to respond to both purinergic and noradrenergic stimulation. Although we observed no gross alterations in the development of ILA processes in FXS, ILAs exhibited enhanced ATP- and NE-evoked Ca²⁺ signaling. Finally, using in vivo two-photon microscopy, we show increased spine elimination and turnover rates in the host dendrites in the vicinity of FXS ILAs. We conclude that Ca²⁺ signaling is altered in FXS ILAs and that the increase in the host dendritic spine plasticity is attributed to the FXS ILAs, indicating the role of astrocytes in the spine dynamics in FXS.

Studies using human brain tissue and human glial chimeric mice have shown differences in Ca²⁺ wave propagation between human and mouse astrocytes [8,28]. The study by Oberheim and coworkers [8] showed that the cell bodies and processes of human astrocytes responded to ATP. The Ca²⁺ imaging studies were performed on slices that were bulk loaded with the Ca²⁺ indicator dye Fluo-4 AM to assess the evoked responses in human astrocytes. However, fluo-4 only allows visualization and quantification of Ca²⁺ signals in somas and proximal processes [29] and is likely to conceal the Ca²⁺ events in the ILA processes. Our study utilized the genetically encoded Ca²⁺ indicator GCaMP6f that overcomes this limitation, thereby revealing the Ca²⁺ signals in the ILA processes. Our study is also the first to show Ca²⁺ signaling in the ILA processes in slices and in vivo. We have previously shown that processes of engrafted ILAs are in putative contact with synapses in the mouse neuropil [15]. Our current studies demonstrating Ca²⁺ along the ILA processes suggest that they could influence synaptic and neuronal activity. However, future studies would have to investigate the consequence of ILA processes activity on neurons.

Rodent astrocytes respond to extracellular ATP and sensory input via elevations of intracellular calcium concentrations [30,31,32]. Astrocytic metabotropic (P2Y) and ionotropic (P2X) purinoreceptors are activated by ATP [1,33]. The transcriptomic study from Zhang et al. (2016) [9] showed that mRNA expression of P2RY1, P2RY12, and P2RY13 were present in human mature astrocytes with a higher expression of P2RY1. While P2Y1 receptors are linked to PLC/IP3/Ca²⁺ signaling cascade, P2Y12 and P2Y13 receptors are linked to G_i proteins and inhibit adenylyl cyclase activity [1]. Interestingly, P2Y12 receptor protein expression has been observed in ILA processes in multiple sclerosis patients [34]. Low mRNA expression of P2RX4 and P2RX7 in human mature astrocytes has also been observed [9]. It is likely that the ATP-evoked Ca²⁺ increases that we observe in the ILA somas and processes are mediated by activation of one or more of these purinoreceptor subtypes that participate in the mobilization of intracellular Ca²⁺ stores.

Astrocyte α1-adrenoreceptor are primary targets for norepinephrine that are coupled into G_q signaling pathways, which trigger increases in Ca²⁺ in rodent astrocytes following startling stimuli or the activation of locus coeruleus (LC) neurons using electrical stimulation [30,32,35,36]. The α2 and β-adrenoreceptors are coupled to G_i and G_s signaling pathways, respectively [37]. Human data indicate ADRA1A, ADRA1B, ADRB1, and ADRB2 mRNA expression in mature astrocytes with a higher expression of ADRB1 and ADRB2 [9]. The norepinephrine-evoked Ca²⁺ increases observed in the ILA somas and processes are likely to be mediated by activation of α1-adrenoreceptors.

Rodent astrocytes have been shown to elicit Ca²⁺ responses with carbachol application that activates G protein-coupled muscarinic receptors [38]. However, human data indicate almost negligible levels of muscarinic cholinergic receptors and low mRNA expression for the nicotinic cholinergic receptors CHRNA4, CHRNA5, and CHRNA7. Consistent with this, we observed no responses in either the ILA somas or processes with carbachol application indicating the absence of functional muscarinic receptors. Put together, our data show that the ILAs in the chimeric mouse have functional purinergic and adrenergic receptors. Ca²⁺ elevation and propagation along the ILA processes is of potential importance as this may facilitate long-distance coordination of intracortical communication.

Long dendritic spines with immature morphologies and higher density have been observed in the postmortem brain tissue of FXS patients [39,40]. Dendritic spine abnormalities reported in the Fmr1 KO mouse model parallel abnormalities reported in FXS patients [41,42]. Fmr1 KO mice have increased rates of dendritic spine formation, elimination, and turnover [19,20,21,43]. Astrocytes have emerged as important regulators of synapse development and have been shown to promote both synapse formation and maturation [2]. Astrocytes can influence both synaptogenesis and synapse maturation through secretion of astrocytic soluble factors and the expression of these proteins has been found to be altered in Fmr1 KO mice, possibly contributing to abnormal neuronal development and altered connectivity observed in FXS [44,45,46,47]. Astrocyte-specific loss of FMRP expression has been reported to result in increased dendritic spine dynamics. These rodent studies have contributed to increasing evidence that astrocytes play essential roles in modulating the function of neurons and neural circuits.

Human astrocytes are more complex than their mouse counterparts, and abnormalities observed in Fmr1 KO astrocytes need to be replicated in human models. Altered structural properties of ILA have been found in other neurodevelopmental disorders. A postmortem study in children with Down Syndrome showed a reduction in the number of ILA processes [48]. Our study in the chimeric mice did not reveal alterations in ILA process length in FXS. Interestingly, dendritic spine density on the dendrites in the vicinity of the engrafted FXS human astrocytes trended higher. Increased dendritic spine elimination and turnover rate were also observed in the FXS group showing that spine dynamics in the host dendrites are altered due to the FXS human astrocytes. Our study is the first to demonstrate altered dendritic spine plasticity using the chimeric mouse model and further highlights the role of astrocytes in the spine dynamics in FXS.

4. Materials and Methods

4.1. Mice

Mice were cared for in accordance with NIH guidelines for laboratory animal welfare. All protocols were approved by the University of Nebraska Medical Center (UNMC) Institutional Animal Care and Use Committee. Rag1 immunodeficient mice (B6.129S7-Rag1^tm1Mom/J, Jackson Laboratory, IMSR Cat# JAX: 002216, RRID: IMSR_JAX: 002216) were bred at the UNMC facility with a 12 h light/dark cycle with food and water available ad libitum.

4.2. Stem Cell Differentiation

The FX11-9u (RRID: CVCL_EJ77), FX11-7 (RRID: CVCL_EJ76) hiPSC lines, and WA01 (RRID: CVCL_9771), H1-FMR1-KO hESC lines were obtained from WiCell. The FX11-9u and FX11-7 hiPSC lines were derived from the same FXS patient. The FX11-9u line retained the FMRP expression due to the mosaicism of CGG repeats during reprogramming [49]. The H1-FMR1-KO (abbreviated as FMR1-KO) was engineered by CRISPR/Cas9 targeting exon 3 of the FMR1 gene in WA01 [50]. The SC176 control line was a gift from Dr. Gary Bassell [51]. The same protocol as previously described [15,16] was used for RFP transduced cells. For the GCaMP6f and mScarlet transduced cells, the embryoid bodies were generated by treatment of hiPSCs with ReLeSR^TM (STEMCELL Technologies; Vancouver, BC, Canada) and passing the clumps though a 100 µm sterile strainer.

4.3. Viral Transduction

NPCs were transfected with CMV-RFP lentivirus (Cellomics Technologies; Halethorpe, MD, USA) or LV-CMV-GCaMP6f-T2A-mScarlet (SignaGen Laboratories; Frederick, MD, USA) as previously described [15] but with the modification that transduced NPCs were expanded prior to differentiation to astrocytes for all cells except those transduced with CMV-RFP.

4.4. Engraftment of Human iPSC-Derived Astrocytes

Engraftment was performed as previously described [15]. Briefly, Rag1⁻ neonatal mice (both males and females) were transplanted on postnatal day 1 with hiPSC-derived astrocytes expressing GCaMP6f and mScarlet or RFP in a random manner. The pups were cryoanesthetized for 4 min and transferred to a neonatal stage (Stoelting) that was cooled to 4 °C during the stereotaxic injections. For cortical labeling, the pups were injected directly through the skin into two sites: AP -1.0 and -2.0, ML ± 1.0 mm, ventral 0.2–0.8 mm, 10,000 cells/µL per site using a Hamilton syringe.

4.5. Brain Slice Preparation

At 6 months following hi-Astrocyte engraftment, anesthetized mice (Avertin, 0.25 mg/g body weight) were transcardially perfused with carbogenated (95% O₂/5% CO₂) N-Methyl-D-glucamine (NMDG) artificial cerebrospinal fluid (aCSF) containing (in mM) 92 NMDG, 2.5 KCl, 1.25 NaH₂PO₄, 30 NaHCO₃, 20 HEPES, 25 glucose, 2 thiourea, 5 Na-ascorbate, 3 Na-pyruvate, 0.5 CaCl₂, and 10 MgSO₄ [52]. Following perfusion, mice were decapitated and their brains quickly removed and immersed in ice-cold carbogenated aCSF. Acute coronal slices containing the frontal cortex were cut to 300 µm using a vibratome. The slices were transferred into a pre-warmed chamber containing carbogenated NMDG aCSF and held for 10 min at 32–34 °C. After this initial recovery period, the slices were transferred into a new holding chamber containing room-temperature recording aCSF containing (in mM) 119 NaCl, 2.5 KCl, 1.25 NaH₂PO₄, 24 NaHCO₃, 12.5 glucose, 2 CaCl₂, and 2 MgSO₄ under constant carbogenation and held in this chamber for at least 1 h before the imaging commenced. For imaging, slices were transferred to the submersion-type chamber and superfused at room temperature (22–24 °C) with recording aCSF saturated with 95% O₂/5% CO₂.

For the Ca²⁺ imaging data shown for 4 months following hi-Astrocyte engraftment, anesthetized mice were transcardially perfused with carbogenated aCSF. Acute coronal slices were cut in carbogenated aCSF containing (in mM) 126 NaCl, 3 KCl, 1.25 NaH₂PO₄, 4 MgSO₄, 2 CaCl₂, 26 NaHCO₃, and 10 dextrose. The slices were transferred into a holding chamber containing room-temperature recording aCSF containing (in mM) 126 NaCl, 3 KCl, 1.25 NaH₂PO₄, 1 MgSO₄, 2 CaCl₂, 26 NaHCO₃, and 10 dextrose.

4.6. Dye Loading in Slices

Following a recovery period, slices were loaded with Fluo-4 AM (5 µM) and 0.6 µL of Pluronic acid F-127 diluted in 1.5 ml of aCSF (containing 1 mM Mg²⁺) saturated with 95% O₂/5% CO₂ for 40 min at room temperature. Following three 10-min washes with aCSF, the slices for imaging were transferred to the submersion-type recording chamber and superfused at room temperature (22–24 °C) with aCSF saturated with 95% O₂/5% CO₂.

4.7. Viral Injection and Cranial Window Implantation

To perform in vivo multiphoton imaging of eGFP expressing mouse neurons in the vicinity of RFP-expressing human astrocytes, we performed viral injection of AAV1.CaMKII.0.4.Cre (1:5000) and AAV1.CAG.FLEX.eGFP into Layer II/III of the cortex and implanted a cranial window in 3.5-month-old chimeric mice [26]. To visualize Ca²⁺ activity in vivo, a cranial window was implanted in 6-month-old chimeric mice expressing GCaMP6f and mScarlet. Mice were injected twice daily with 5 mg/kg enrofloxacin for 6 days and 5 mg/kg carprofen daily for 20 days. Mice were allowed three weeks to recover from the surgery prior to imaging.

4.8. Multiphoton Imaging

4.8.1. Slice Imaging and Pharmacology

Time-lapse imaging was performed with a two-photon microscope (Moving Objective Microscope (MOM), Sutter; Novato, CA, USA) attached to a Ti: sapphire laser (Chameleon Vision II, Coherent; Santa Clara, CA, USA) using a 25× water immersion objective (1.05 NA, Nikon; Minato City, Tokyo, Japan) and equipped with a resonant scanner and a piezo stage (nPFocus400, nPoint; Middleton, WI, USA). Excitation wavelength was tuned to 920 nm with 10–12 mW power, as measured at the back aperture. Two-channel detection of emission wavelength was achieved by using a 670 nm dichroic mirror and two external photomultiplier tubes (GaAsP). A 535/50 bandpass filter was used to detect GCaMP6f emission wavelength, and a 610/75 bandpass filter was used to detect mScarlet or RFP. For imaging, we used ScanImage software v.2020.1.0 (MBF Biosciences, Williston, VT, USA) written in MATLAB v. R2024b (The MathWorks; Natick, MA, USA) [53]. Images of ILA were acquired at a resolution of (0.3 µm/pixel), a 20 µm volume at a step size of 1 µm and a frame rate of 1 Hz. Time-lapse imaging was performed for a period of 5 min. For each slice, two minutes of spontaneous Ca²⁺ activity was recorded first, followed by agonist application for 2 min. Agonists were added to the perfusing ACSF at the following final concentrations: 100 µM ATP, 50 µM norepinephrine (NE), and 50 µM carbachol. The addition of ATP had no effect on the pH of the ACSF. Slices were exposed only to a single agonist.

4.8.2. In Vivo Imaging

Dendritic spine imaging: On the imaging day, 4-month-old chimeric mice were anesthetized with a ketamine/dexdormitor mixture (100 mg/mL and 0.5 mg/mL, respectively, dosage 2.5 mL/kg). Excitation power measured at the back aperture of the objective was typically around 20 mW and was adjusted to achieve near identical levels of fluorescence for each imaged region. Each optical section was collected at 512 × 512 pixels (0.186 µm/pixel). During an imaging session, 5 to 10 ROIs per animal were selected along with the apical dendritic tufts of eGFP expressing neurons within the vicinity of the RFP expressing human astrocytes in the cortex. Each ROI consisted of a stack of images (50–80 optical sections, separated axially by 1 µm). The coordinates of each ROI were recorded using the XYZ motor on the MOM for subsequent imaging days. After imaging, mice were revived from anesthesia with antisedan (atipamezole hydrochloride 5.0 mg/mL). Dendritic spines were imaged twice, at a four-day interval.

In vivo calcium imaging: Mobile home cage (MHC, Neurotar; Helsinki, Finland) was used for imaging in awake mice where a head-fixed mouse can move around in an air-lifted MHC that features a flat floor and tangible walls and explore the environment under stress-free conditions. Prior to the imaging experiment, the mouse was habituated to the MHC by gradually increasing the duration of the habituation sessions every day and acclimating the mouse to the sounds of the laser scanning mirrors [26]. The habituation phase started two weeks after the cranial window was implanted. Time-lapse imaging was performed every 2 s for a period of 5 min. Each optical section was collected at 512 × 512 pixels, 0.37 μm/pixel and a 20 µm volume was acquired at a step size of 1 µm.

4.9. Image Analysis

Process length analysis was performed as previously described [15]. Briefly, in each section multiple 5-pixel broad straight lines, 150–200 µm apart, were drawn from the pial surface to the deeper layer. For each line, the plot profile function was used to estimate the pixel intensity values along the line. The location of the last peak of RFP fluorescence was used as a measure of the extent to which processes traversed across the cortex and averaged per section. The values from control FX11-9u line were previously published [15]. Analysis was performed while blinded to engrafted line identity.

4.9.1. Analysis of Astrocytic Ca²⁺ Events

The images acquired as a 20-micron volume were maximum intensity z-projected in 5-micron volumes for further analysis. For proportion of active ILA processes up to 10 processes per field of view were scored as to whether they were responsive to an agonist. Event-based analysis of astrocyte Ca²⁺ image events was performed using Astrocyte Quantitative Analysis (AQuA v.2020) software (Guoqiang Yu lab, Virginia Tech; Blacksburg, VA, USA) [54] only on a subset of well-isolated processes (1–3 per field of view). The AQuA-detected events were categorized based on the area of the Ca²⁺ events with 3 < 10 µm² (small) and >10 µm² (large). Not all processes displayed both small- and large-sized events. For analysis of astrocytic Ca²⁺ activity in the soma, a region of interest was outlined around the soma. The Ca²⁺ events were categorized into spontaneous and agonist-evoked events. The peak amplitude (dF/F), area under the curve, duration (at half width), rise time (10–90%), and decay time (90–10%) of the events were analyzed. Clampfit v10.6 (pCLAMP, Molecular Devices; San Jose, CA, USA) software was used to detect and measure the parameters for the soma. Analysis was performed while blinded to engrafted line identity.

4.9.2. Analysis of Spine Plasticity

The analysis of spine plasticity was performed on ImageJ v.1.54 software (Rashband, W. NIMH, NIH; Bethesda, MD, USA). For each image, dendrites with at least 25 µm-length, as well as those located within 5 focal planes and 20 µm XY coordinates from the RFP-expressing astrocytes were selected for further analysis. All spines on the selected dendrites were counted and tracked over time to identify newly formed, eliminated, and stabilized spines. Dendritic filopodia were distinguished as long dendritic protrusions with no head and were excluded from analysis (<3% in both genotypes). Dendritic spines were analyzed by scrolling through individual z-planes within a stack. Spines were categorized as stable if they were present in the previous image as well as the one being analyzed, eliminated if they appeared in the previous image but not in the image being analyzed, and newly formed as they appeared in the image being analyzed but not in the previous image. The percentage of spine formation and elimination was calculated using the number of newly formed or eliminated spines to the total number of spines analyzed in the image, respectively. Turnover rate (TOR) was calculated as the ratio of the sum of newly formed and eliminated spines to twice the total number of spines at baseline [20]. Analysis was performed on raw unprocessed images. For presentation purposes, images were despeckled and put through maximum intensity projection of 3–15 planes of focus. Analysis was performed while blinded to engrafted line identity.

4.10. Tissue Preparation and Confocal Imaging

Mice were deeply anesthetized with TribromoEthanol (Avertin, 400 mg/kg i.p.) and transcardially perfused with 4% paraformaldehyde in a phosphate buffer (0.1 M) at 3 and 9 months post-engraftment. The brain was dissected, post-fixed overnight, and 100 µm sagittal sections were cut on a vibratome in PBS. Confocal imaging of tissue sections was performed on a Nikon A1R upright microscope and images were acquired using a 20× (0.75 NA) objective. Images were collected at 512 × 512 pixels (with a pixel size of 1.24 μm) and a step size of 1 μm with 561 nm laser.

4.11. Statistical Analysis

Data were analyzed using GraphPad Prism v.10.5.0 (Graphpad Software, LLC; San Diego, CA, USA). Outliers were removed using the ROUT approach (Q = 1%). Normal distribution was tested using the Shapiro–Wilk test. Data are reported as mean ± SEM or median ± interquartile range. On scatter plots, medians are denoted by a thin line and means by a thick line. The following tests were conducted as appropriated: unpaired t-test, Mann–Whitney test, and two-way ANOVA. p values were adjusted for multiple comparisons. For proportion of responding ILA processes, a Chi-Square test was used.