This protocol describes a chronic cranial window implantation technique that can be used for longitudinal imaging of neuro-glio-vascular structures, interactions, and function in both healthy and diseased conditions. It serves as a complementary alternative to the transcranial imaging approach that, while often preferred, possesses some critical limitations.
The central nervous system (CNS) is regulated by a complex interplay of neuronal, glial, stromal, and vascular cells that facilitate its proper function. Although studying these cells in isolation in vitro or together ex vivo provides useful physiological information; salient features of neural cell physiology will be missed in such contexts. Therefore, there is a need for studying neural cells in their native in vivo environment. The protocol detailed here describes repetitive in vivo two-photon imaging of neural cells in the rodent cortex as a tool to visualize and study specific cells over extended periods of time from hours to months. We describe in detail the use of the grossly stable brain vasculature as a coarse map or fluorescently labeled dendrites as a fine map of select brain regions of interest. Using these maps as a visual key, we show how neural cells can be precisely relocated for subsequent repetitive in vivo imaging. Using examples of in vivo imaging of fluorescently-labeled microglia, neurons, and NG2+ cells, this protocol demonstrates the ability of this technique to allow repetitive visualization of cellular dynamics in the same brain location over extended time periods, that can further aid in understanding the structural and functional responses of these cells in normal physiology or following pathological insults. Where necessary, this approach can be coupled to functional imaging of neural cells, e.g., with calcium imaging. This approach is especially a powerful technique to visualize the physical interaction between different cell types of the CNS in vivo when genetic mouse models or specific dyes with distinct fluorescent tags to label the cells of interest are available.
The central nervous system (CNS) is governed by a complex interplay of interactions between various resident cell types including neurons, glia and vessel-associated cells. Traditionally, neural cells were studied in isolated, co-cultured1,2,3,4,5 (in vitro) or excised brain tissue (ex vivo)6,7,8,9,10 contexts. However, there is need to further understand neural cell behavior and interactions in the native environment of the intact brain in vivo. In this protocol, we describe a method to map in vivo regions of interest and precisely re-image those regions in future imaging sessions to track the complex interactions between the various CNS cell types over extended periods of time.
The development of in vivo imaging approaches has provided significant gains for the proper understanding of neural function11,12,13,14,15. Specifically, these approaches provide several advantages over traditional in vitro and ex vivo approaches. First, in vivo imaging systems have physiologically relevant cell and tissue components such as the vasculature with the full repertoire of cellular interactions to provide a complete understanding of neural network physiology. Second, recent findings suggest that when removed from their native environment, certain neural cells (such as microglia) lose important features of their identity and thus physiology16,17 which can be preserved in the in vivo setting. Third, in vivo imaging systems provide the opportunity for stable longitudinal investigations of weeks to months to study CNS cellular interactions. Finally, given the growing evidence for contributions from the peripheral immune system18,19 and the microbiome20,21 in CNS physiology, in vivo systems provide a platform to interrogate such contributions and effects on CNS cells. Thus, approaches that employ longitudinal in vivo imaging to study neuro-immune physiology and interactions in healthy, injured, and diseased states are a great complementary addition to traditional approaches.
In this protocol, we describe a reliable approach to image different cell types in the brain including microglia, neurons and NG2+ cells as examples. Two approaches to visualize neural cells in vivo have been developed: the thinned skull approach and the open skull with a cranial window approach. Although thinned skull approaches are in use and are preferred because they overcome some of the disadvantages of the open skull approach such as glial cell activation, higher-than-physiological spine dynamics and the use of anti-inflammatory agents22,23,24,25, thinned skull approaches also show a few critical drawbacks. First, the thinning procedure is a very delicate procedure that many researchers find difficult to perfect especially when re-thinning is necessary. This is the case because it is often difficult for experimenters to ascertain that they have thinned the skull to a ~20 µm depth. Second, for adequate comparisons between mice, thinning would need to be identical and a variety of thinning success between imaging sessions or mice could complicate visualization of neural structures. Third, when employed for longitudinal imaging, animals with thinned skulls can only be used for a limited number of sessions when re-thinning of the skull is employed. Forth, since some of the bone tissue still remains, clarity in depth of imaging could be compromised from the thinned skull approach allowing for great visualization of more superficial but not as much with deeper regions. In the light of this, deeper brain structures such as the hippocampus, cannot be successfully imaged with the thinned skull approach. These considerations raise the need for alternative and complementary approaches that could overcome these concerns.
Alternative to the thinned skull approach, the open skull window implantation approach uses a procedure in which the skull is replaced with an optically clear glass coverslip. This allows for a near-unlimited number of imaging sessions. Moreover, given the replacement of the skull with the glass coverslip, this method allows for a clear viewing window of fluorescently tagged brain cells for extensive periods of times from hours to months and, therefore, can be employed to study cell activity and interactions that are relevant for physiology, aging and pathology.
Overall, we detail steps that can be followed to do implant chronic cranial windows through a stereotaxic craniotomy that enables in vivo imaging of brain regions of interest. We also describe how the grossly stable brain vasculature or the fluorescently labeled dendrites could be used to generate a coarse or a fine map, respectively of the brain regions of interest. This approach can then be used for repeated imaging over several sessions. The importance of this technique, therefore, lies in its ability to image the long-term changes or stasis in brain elements including the arrangement, morphology, and interactions of the different cellular types.
All steps are in accordance with the guidelines set and approved by the Institutional Animal Care and Use Committee of the University of Virginia.
1. Mouse preparation for cranial window implantation
NOTE: Various transgenic mouse lines with florescent tags are suitable for imaging.
2. Mouse cranial window implantation surgery
3. Post-surgery care
4. Two-photon brain mapping for initial imaging
5. Two-photon imaging and re-imaging
To visualize microglial dynamics in vivo, double transgenic CX3CR1GFP/+:Thy1YFP mice were used. The Thy1-YFP H line is used as opposed to the Thy1-GFP M line to avoid florescence overlap of microglia (GFP) and neurons (YFP). Alternative approaches could use a reporter line in which microglia are labeled with e.g., tdTomato and then the Thy1-GFP M line can be used. A drawback of the H line is that YFP labels a lot of neurons and the label increases with increasing age (personal observation). The M line exhibits sparse labeling of neurons. Between 2 – 4 weeks of the window implantation surgery, microglial dynamics can be followed by repeated imaging. Large blood vessels are used to localize specific regions and then the YFP-labeled dendrites are used for fine mapping of brain regions. With this approach, specific dendrites can be used as stable landmarks for the fine mapping of brain regions (Figure 1, arrows in Figure 1b). While dendrites are stable, some microglia move daily (Figure 1c).
In addition, this approach is sufficient for weekly longitudinal imaging in the long-term. Thus, single transgenic CX3CR1GFP/+ mice were used to follow microglia coupled with intraperitoneal injections of Rhodamine B to label the vasculature during each imaging session for up to 8 weeks (Figure 2). Alternatively, as discussed above, Thy1 mice could be used for longitudinal fine mapping. When weekly imaging is performed, the vasculature is noted to be stably fixed, but microglia can be seen to be dynamic as shown in three specific regions of interest (ROI, dashed circles) in Figure 2. In the top ROI, microglia begin to enter the ROI by the 4th week of imaging and continue through the 8th week of imaging. In the middle ROI with a bifurcated vessel, a microglial cell emerges around the lower vessel in the 3rd week, is lost on the 6th week and another microglia emerges on the upper blood vessel in the 7th week and is maintained into the 8th week. Finally, in the bottom ROI, a microglial cell in maintained through the 6th week and lost in the 7th and 8th weeks of imaging. These results indicate dynamic changes in the microglial positional network over weeks to months.
This approach can also be used to investigate cellular dynamics following acute injury or during pathological disease progression. Single transgenic CX3CR1GFP/+ mice were used to follow microglia coupled with intraperitoneal injections of Rhodamine B to label the vasculature before (data not shown) and following severe seizures induced by kainic acid (Figure 3). Following seizures, the vascular bed structure is maintained without overt perturbations (Figure 3a). However, the microglial cellular network and positional landscape is transiently changed (some cells are “gained” and others are “lost” in the field of view) with greater changes within 24-48 h of seizures that is restored to normal by 72 h (Figure 3b) as we previously reported32.
Finally, this approach can also be used to investigate cell-cell interactions or compare dynamics between neural cell types. Double transgenic CX3CR1GFP/+:NG2dsRed/+ mice were used to track microglia and NG2+ cells in vivo. Without labelling the vasculature, microglia and NG2 cells can be identified (Figure 4a). NG2 is a proteoglycan that labels both vessel-associated pericytes and oligodendrocyte precursor cells (OPCs)33,34. Pericytes typically have elongated processes that follow along the vascular wall (presumptively identified with arrowheads in Figure 4a) and OPCs typically show larger cell bodies that reside in the brain parenchyma away from the vasculature (presumptively identified with arrows in Figure 4a). To adequately distinguish pericytes and OPCs, the vasculature is labeled with Rhodamine B. The brighter florescence of NG2+-vessel associated cells (pericytes, arrowheads) can be distinguished from the fainter florescence of luminal Rhodamine despite similar excitation by two photon imaging (Figure 4b,c). Daily imaging shows that pericytes are stably positioned, while OPCs (asterisks in Figure 4b) and microglia (circles in Figure 4b) are dynamic consistent with previous reports32,35,36.
Figure 1: Daily imaging of microglia using fine mapping with neuronal dendrites in double transgenic CX3CR1GFP/+:Thy1YFP mice. (a) Representative two-photon image of microglia (green) and dendrites (red) from a double transgenic CX3CR1GFP/+:Thy1YFP mouse. (b-c), Daily images of boxed region in (a) showing repeatedly imaged dendrites (arrows in b) and dendrites with microglia (c). While dendritic structures were positionally stable, some microglia were noted to translocate from their original position in subsequent days. Such cells were identified with a number (1, 2 or 3). On the previous day, their position was noted with a white asterisk and on a subsequent day, their position was noted with a yellow asterisk. Please click here to view a larger version of this figure.
Figure 2: Long-term weekly imaging of microglia in CX3CR1GFP/+ mice for several months. (a-h), Representative two-photon images of microglia (green) from a CX3CR1GFP/+ mouse during repeated imaging using acutely labeled vasculature (red, Rhodamine, 2 mg/mL, i.p.) as a coarse landmark to track the microglial network for up to 8 weeks. The vasculature was structurally stable through the imaging period. Three small regions with the vasculature (dashed circles) were highlighted to indicate the movement of microglial somata into (top two dashed circles) or out of (bottom circle) those regions. Please click here to view a larger version of this figure.
Figure 3: Long-term daily imaging of microglia in CX3CR1GFP/+ mice following seizures. (a-b), Representative two-photon images during daily imaging of the same field of view of a specific brain region with microglia (green) and acute labeling of the vasculature (red, Rhodamine, 2 mg/mL, i.p.). Imaging begins after induction of severe seizures using a chemoconvulsive agent, kainic acid. The vascular structure was maintained as individual vascular segments (arrows) can be identified through time (a). However, the microglial network dynamics was increased during the first two days following the seizures and returns to normal levels by the third day (b). Please click here to view a larger version of this figure.
Figure 4: Daily imaging of microglia and NG2+ cells in CX3CR1GFP/+:NG2dsRed/+ mice. (a) A representative two-photon image of microglia (green) and NG2+ cells (red) in vivo. The unlabeled vasculature fails to distinguish NG2 cells (pericytes, presumptively identified with arrowheads) and NG2 cells not associated with the vasculature (oligodendrocyte precursor cells or OPCs, presumptively identified with arrows). (b) Representative two-photon images of microglia (green) and NG2+ cells (red) in consecutive days of imaging with the vasculature labeled with Rhodamine. Pericytes (arrowheads) are stationary while OPCs are dynamic (white to yellow asterisks). Microglia are also dynamic (circles: dashed circles represent a position without microglia and filled circle represents a position with a corresponding microglia). Please click here to view a larger version of this figure.
The advent of in vivo two-photon imaging has opened opportunities to explore the plethora of cellular interactions and dynamics that occur in the healthy brain. Initial studies focused on using the open skull craniotomy approach to image neuronal dendrites by both acute and chronic imaging37,38. This can also be used to elucidate neuroimmune interactions in the brain. This protocol describes a method for the reliable imaging of fluorescently tagged cells (especially microglia, the resident immune cell of the brain) for extended periods of time in the short or long term. The use of dye-labeled vascular and / or fluorescently tagged dendrites is detailed for coarse or fine mapping of the brain regions of interest to allow repeated, reliable imaging of cells. Although the Thy1YFP line is suggested for use for fine brain mapping, alternative approaches could use other techniques or mouse lines for labeling select neuronal populations such as in utero electroporation39,40, early postnatal AAV injections41 or the use of TRAP mice42. Furthermore, although cortical imaging was the focus of this discussion, this approach can be adapted to visualize deep brain structures in the long-term as well43.
For this protocol, the surgery procedure on each mouse can be completed in 30-60 min from the initiation of anesthesia until the recovery from anesthesia. During the surgery, the skull is carefully removed and replaced with a sterile coverglass which is implanted for long-term imaging after at least two weeks. Mortality is extremely rare (less than 5%) in wildtype mice, though mice with clotting problems, such as P2Y12R knockout mice, show higher mortality and surgery failure. In such mice, bleeding may persist for longer periods of time and mice may die within the first 48 h of the craniotomy presumably due to complications from internal bleeding. Mice with implanted windows from this protocol have not been noticed to show any signs of infection and the protocol can be reliably used to generate clear windows for long-term imaging in 50-80% of mice.
Alternative to the current chronic window implantation approach, the thin skull approach exists to visualize brain cells in the intact brain repeatedly. Several studies have highlighted the value and even priority of choice of the thin skull approach over the window implantation approach22,23,24,25. The promise of that approach should not be ignored as, when done properly, it counteracts several salient limitations or disadvantages of the current approach including the lack of activation of glial cells, the low turnover of spines which is more physiological, and the lack of need for the use of anti-inflammatory agents which could also affect brain physiology. In selecting an approach for specific research questions, these serious limitations should be considered before choosing the approach detailed in this protocol.
However, the appeal of this approach is four-fold. First, the cranial window implantation approach is attractive because of the ease of mastery of this procedure relative to the thin skull procedure. Appropriate skull thinning cannot always be mastered by experimenters and if not done well can result in glial activation limiting its appeal. Second, this cranial window implantation approach gives powerful depth clarity of brain structures as the brain is imaged through an optically clear window. The window available for imaging is also usually much larger than that used in the thin skull approach allowing access to a larger volume of tissue for analysis. Third, like depth clarity, this approach allows for a uniform clarity through the window since the glass coverslip is uniformly thin and clear. This facilitates comparisons between sessions and between animals. Special expertise is required for the thin skull technique to ensure even clarity across the window during repeated sessions and between animals. Finally, this approach offers much flexibility in the frequency of imaging from hours, to days to weeks to months and even years. For the thin skull approach, a maximum of five repeats has been suggested24.
Future applications of this approach are many. First, applications could involve elucidating novel neuro-glio-vascular interactions in the brain in both normal physiology and pathology. Second, although resident cells are discussed in this procedure, the approach can be used to study the dynamics and interactions of infiltrating immune cells as occurs e.g., during acute injury, chronic brain infection, and/or neurodegenerative conditions as long as mice with the respective fluorescently tagged cells are available. Finally, this approach has been discussed mainly in the context of structural studies of brain cells. However, with the advent of functional imaging e.g. using calcium44,45,46 or voltage imaging techniques47,48, this approach can be used for functional imaging over time in health and disease.
The authors have nothing to disclose.
We thank members of the Eyo lab for discussing the ideas presented in this manuscript. We thank Dr. Justin Rustenhoven from the Kipnis Lab at the University of Virginia for the gift of NG2DsRed mice33. This work is supported by funding from the National Institute of Neurological Disorders and Stroke of the National Institute of Health to U.B.E (K22 NS104392).
Coverglass (3mm) | Warner Instruments | 64-0726 | |
Cyanoacrylate glue (Krazy Glue) | Amazon | https://www.amazon.com/Krazy-Glue-Original-Purpose-Instant/dp/B07GSF31WZ/ref=sr_1_2?keywords=krazy+glue&qid=1583856837&s=pet-supplies&sr=8-2 | |
Demi Ultra LED Curing Light System | Dental Health Products, Inc | 910860-1 | |
Dental Drill | Osada: www.osadausa.edu | EXL-M40 | |
Drill Bit | Fine Science Tools | #19008-07 | |
Eye Ointment | Henry Schien | 1338333 | |
iBond Total Etch (Primer glue) | Chase Dental Supply (Heraeus Kulzer) | 66040094 | |
Rhodamine B | Millipore Sigma | 81-88-9 (R6626) | |
Tetris Evoflow glue (Final glue) | Top Dent (Ivoclar Vivadent) | #595956 | |
Wahl Brav Mini+ | Amazon | https://www.amazon.com/Wahl-Professional-Animal-BravMini-41590-0438/dp/B00IN24ILE/ref=asc_df_B00IN24ILE/?tag=hyprod-20&linkCode=df0&hvadid=167141013968&hvpos=&hvnetw=g&hvrand=12368793083893626704&hvpone=&hvptwo=&hvqmt=&hvdev=c&hvdvcmdl=&hvlocint=&hvlocphy=9008337&hvtargid=pla-332197544154&psc=1 |