2.1.

Animals

Tg(Id3-EGFP)FS137 Gsat mice were obtained from Mutant Mouse Regional Resource Centers (MMRRC) University of California, Davis.²⁰ Mice of different ages, including early postnatal development—P7; adolescence—P28; and early adulthood—P61, were used to characterize the expression of EGFP. GLT-1 eGFP BAC promoter reporter mice²¹ were used as adults ($>\mathrm{P}90$). The animal and experimental protocols were approved by the University of Rochester University Committee on Animal Resources (UCAR) in accordance with the Public Health Service (PHS) policy on humane care and use of laboratory animals and conformed to the National Institute of Health Guidelines.

2.2.

Immunohistochemistry

Mice were deeply anesthetized with sodium pentobarbital ($80\mathrm{mg}/\mathrm{kg}$, i.p.) and perfused transcardially with 100-mM phosphate buffered saline (PBS) and then with 4% paraformaldehyde (PFA) in 100-mM phosphate buffer. After postfixation in 4% PFA at 4°C, the brains were cryoprotected in 10%, 20%, and 30% sucrose and flash frozen. The brains were coronally sectioned at $50\mu \mathrm{m}$ on a freezing microtome (HM 430 Microtome; MICROM International GmbH, Walldorf, Germany). The tissue was cut into cryoprotectant and frozen at $-20°\mathrm{C}$. For immunofluorescence staining, the tissue was removed from the cryprotectant and washed three times for 15 min each in 0.1 M PBS. For analysis of EGFP expression, six animals per age group and three sections of tissue per area were analyzed. For immunohistological analysis, sections were placed in a solution containing 3% ${\mathrm{H}}_2{\mathrm{O}}_2$ and 0.1 M PBS for 20 min. Sections were washed three times for 15 min in 0.1 M PBS. Sections were then blocked for 1 h in a solution containing 5% normal donkey serum, and 0.3% Triton-X-100 in 0.1 M PBS. Sections were then washed again in 0.1 M PBS and placed in the primary antibodies: rabbit anti-Iba1, (1:500, Wako Chemicals USA, Inc., Richmond, Virginia), mouse anti-neuronal nuclei (NeuN, 1:500, Millipore, Billerica, Massachusetts), rabbit anti-glial fibrillary acidic protein (GFAP, $1:2000$, DAKO), rabbit anti-GFP ($1:3000$, Invitrogen, Grand Island, New York) at 4°C, shaking for 48 h. Sections were washed and incubated in secondary antibodies (AlexaFluor 594 Donkey anti-rabbit IgG, $1:250$; AlexaFluor 594 Donkey anti-mouse IgG, $1:250$; AlexaFluor 594 Donkey anti-mouse IgG, $1:250$; Alexa 594 Donkey anti-rabbit $1:500$, Molecular Probes, Carlsbad, California) for 4 h. Sections were rinsed in 0.1 M PBS, mounted on slides, and coverslipped with ProLong Gold anti-fade (Molecular Probes, Carlsbad, California) then edges were sealed and slides were protected from light until analysis.

2.3.

Image Acquisition and Analysis

The mouse brain atlas was used to locate brain areas to be analyzed.²² For analysis of EGFP expression, images were taken using epifluorescence on a Olympus BX51 epifluorescence microscope (Olympus, Center Valley, Pennsylvania) using a $10\times 0.3\mathrm{NA}$ objective and a $20\times 0.5\mathrm{NA}$ objective (Olympus). Digital images were acquired with a Spot Insight Color camera (Diagnostic Instruments, Sterling Heights, Michigan) and Image Pro software (Media Cybernetics, Bethesda, Maryland), and acquisition parameters were kept constant for imaging of all sections. An observer blinded to age analyzed the expression of EGFP in astrocytes and blood vessels with a scoring system where 0 corresponded to no staining and 3 corresponded to intense staining. Scores from all animals were averaged and rounded to the nearest number. The density of labeled structures was scored with the same system. Lastly, the appearance of astrocytes was scored as N (for net—many astrocytes grouped together) and S (single—individual astrocytes are labeled with clear visualization of the process arbor). For observation of co-localization of EGFP with different immunohistological markers, images were collected on a Zeiss LSM 500 confocal microscope (Thornwood, New York) using a $40\times 1.2\mathrm{NA}$ lens and a $100\times 1.46\mathrm{NA}$ lens (Zeiss). Co-localization was observed qualitatively and no quantification was attempted.

2.4.

In Vivo Two-Photon Imaging

For two-photon imaging, mice were anesthetized with avertin ($16\mu \mathrm{g}/\mathrm{g}$ of body weight; i.p.); the skull was exposed and cleaned of membranes. The skull was then dried and glued to a thin metal plate using Loctite 454 gel glue (Prism, Rocky Hill, Connecticut). The glue was dried using an accelerator (Zipkicker, Pacer Technology, Rancho Cucamonga, California), which was pipetted onto the skull. Care was taken not to allow the accelerator to touch the mouse’s eyes. Primary visual cortex (V1) was identified according to stereological coordinates. The skull above the imaged area was thinned with a dental drill or a small craniotomy was made and sealed with agarose and a coverslip.²³ The procedure was frequently interrupted to apply cold saline to the skull in order to prevent brain injury and astrocytic activation. During surgery and imaging, the animal’s temperature was kept constant with a heating pad and the anesthesia was maintained with periodic administration of avertin. A custom-made two-photon scanning microscope²⁴ was employed, using a wavelength of 920 nm and a $20\times 0.95\mathrm{NA}$ objective lens (Olympus, Melville, New York) at $10\times$ digital zoom. Images were acquired as $z$ stacks with a $1–5\text{-}\mu \mathrm{m}$ step size. Time-lapse images were acquired as $z$-stacks every 5 min.

4.1.

Id3 Expression in the Postnatal Brain

Although our main goal was to characterize this mouse line for the in vivo imaging of astrocytes, our results also give us insight into the postnatal expression of Id3. Id3 has known roles in early gestational development, where it is critical for promoting angiogenesis and neurogenesis.¹⁹ Much less is known about its function in the postnatal brain, although it has been implicated in tumor growth,¹⁹ immune cell differentiation,²⁵ wound healing,²⁶ inflammation,²⁷ and blood vessel repair.²⁸ Our findings imply that Id3 expression is developmentally regulated in a cell-type specific manner within different brain regions postnatally, and is restricted to astrocytes and blood vessel wall associated cells. This suggests that Id3 has important roles in the brain vasculature throughout life and is important in the regulation of gene expression in a small number of postnatal astrocytes. In these studies, we focused our attention on astrocytes, therefore, we cannot definitely say which cells in the vascular walls express Id3, although the sparse pattern of astrocyte labeling makes it unlikely that Id3 labeling of the vasculature results from its expression in astrocytic endfeet. Id3 has been shown to be expressed in endothelial cells during development and in tumor vasculature, although its expression is thought to be limited in the adult brain. It is important to note that we could not verify our pattern of Id3 expression in wild-type mice due to the lack of specific tools to track the endogenous Id3 protein. This characterization will be important in future studies of Id3 expression in postnatal animals.

4.2.

Imaging Astrocytes and Blood Vessels in the Intact Brain

Whether the BAC transgenic mouse faithfully follows the endogenous Id3 expression pattern is less relevant for our main goal of characterizing EGFP expression in astrocytes in these mice. Our results show that EGFP is expressed throughout early postnatal development and adulthood and labels a small proportion of the astrocytic population. This is reminiscent of the thy-1 XFP mice,²⁹ which are very popular for in vivo imaging studies of neuronal architecture. The sparse labeling of astrocytes allows for easier imaging of fine processes—in fact such structure is lost in areas where many astrocytes are labeled such as in layer 1 of the cerebral cortex in BAC Id3 EGFP mice (Fig. 5), and in GLT-1 mice in which the majority of astrocytes express EGFP (Fig. 9). It is unclear what governs astrocytic expression of EGFP in the BAC Id3 EGFP mouse line. This labeling could be stochastic—where each astrocyte may or may not express EGFP relatively randomly, or labeling may delineate a specific subclass or a small number of subclasses of astrocytes. If the latter is true, this mouse line may allow the imaging of a new class of astrocytes defined by Id3 expression. In fact, most immunohistological markers of astrocytes label only a small fraction of brain astrocytes suggesting that specific functional subtypes, defined by their expression patterns, may exist in the brain. Unfortunately, astrocyte heterogeneity is still a very poorly understood topic,³⁰ making it all the more important to define and explore different classes of astrocytic cells.

There are many tools for labeling astrocytes in vivo because of the importance of studying these cells in their native milieu. Of the organic dyes, sulforhodamine 101 (SR101)³¹ has proven to be the most reliable for identifying astrocyte, although its application often requires the opening of the skull and is, therefore, generally only useful for acute experiments (see Refs. 32 and 33). It has been used extensively and can label astrocytic processes.³³ In the majority of studies, however, labeling is largely limited to the soma. SR101 use is further limited by its uptake being age-dependent and brain-region specific,^34,35 and SR101 has also been shown to interfere with plastic processes.³⁶ Astrocytes have also been labeled using several virus labeling strategies,^15,16,37,38 although studies of astrocytic processes in viral labeled astrocytes have been limited to ex vivo preparations.^15,16 The downside of the virus labeling approach is that injection of the virus into the brain can cause damage and inflammation which may alter the course of the experiment. Other transgenic approaches are also available. Most common are the GFAP-EGFP mice, which label only those astrocytes with the highest GFAP expression.^39,40 These mice have been used to image astrocytic fine process dynamics in slice cultures in the cerebellum⁴¹ and astrocytic responses to injury in vivo.^12,13 The Aldh1l1-EGFP and mice label the entire astrocyte population.^13,20,42 In these mice, individual processes are not apparent either due to the dense labeling or to the intensity of EGFP expression. Similarly, in our study, the BAC GLT-1 EGFP mouse²¹ allowed the visualization of a large population of astrocytes in the superficial layers of visual cortex in vivo without providing a clear image of individual fine processes (Fig. 9). In contrast, an astrocytic line that labels the microtubule cytoskeleton has been used to obtain beautiful images of astrocytic processes in vivo,¹⁸ however, this line may miss the structure and dynamics of areas supported by other cytoskeletal proteins. Recently, ${\mathrm{GLAST}}^{\mathrm{CreERT}2}$ knock-in mice⁴³ crossed with an inducible reporter line have been used to image astrocytes in vivo.¹³ A subset of astrocytes, including the astrocytic process arbor, was clearly visible, but the authors did not attempt to image fine morphology. Such a conditional approach could be adapted to the many astrocytic driver Cre mice available and may prove very useful for in vivo imaging of fine astrocytic processes in the future.

4.3.

Dynamics of Astrocytic Processes

Astrocytes interact with synaptic structures and can modulate synaptic transmission and plasticity. Because the coverage of synapses by astrocytic processes is sensitive both to physiological and pathological stimuli,^44,45 it has been proposed that astrocytic processes can be highly dynamic. In organotypic slice cultures, virally labeled astrocytic processes were shown to be more highly motile (on the order of minutes) than their postsynaptic partners.¹⁶ Such dynamic synaptic interactions could serve to stabilize specific synapses within a circuit.^15,17 While the activity-dependent motility of glial processes has been demonstrated in vivo in developing brains of unanesthetized xenopus tadpoles,⁴⁶ to date there is little evidence to support the high motility of fine astrocytic processes in adults in vivo, especially in mammalian systems. This is in part due to the dearth of labeling tools that allow for the visualization of the very small tips of astrocytic processes within the intact brain. The Id3 BAC transgenic mouse allowed us to image the dynamics of fine processes of adult astrocytes in vivo. We did not observe any motility of fine astrocytic processes on the timescale of tens of minutes, suggesting that astrocytes in the adult intact brain of anesthetized animals may not be motile. This agrees with in vivo imaging of microtubule structures in astrocytic processes, which was also described to be stable on these timescales,¹⁸ and with our time-lapse studies of the dynamics of very superficial fine processes in the BAC GLT-1 EGFP mouse. While we cannot rule out that processes that fall under our detection limit may be motile, it is interesting to note that we could visualize processes that were $\sim 1\mu \mathrm{m}$ in size, similar to the size of motile processes in previous in vitro studies of developing systems.¹⁶ In contrast, it is likely that adult astrocytic processes are motile on a slower scale, as observed in response to injury.¹³ Such slow motility of astrocytic processes could also be mediated by water fluxes in astrocytes, as astrocytes have been shown to swell in both physiological and pathological conditions,^12,47 regulating extracellular space and intercellular signaling. Recently, the swelling of astrocytes has been shown to be regulated by the sleep wake cycle, allowing the clearance of brain metabolites during sleep.⁴⁷ Such subtle changes in astrocytes between different physiological states may regulate astrocytic actions at synapses in the adult more than rapid motility of processes. On the other hand, astrocytic process motility may be differentially regulated in awake and anesthetized animals as has been shown for other astrocytic functions.^9,48 More investigation of the dynamics of astrocytic process arbors in vivo in awake animals, both during development and adulthood, using new labeling strategies such as the Id3 BAC transgenic mouse, will be needed to fully understand the role of rapid motility of astrocytes under physiological conditions.