Modified Roller Tube Method for Precisely Localized and Repetitive Intermittent Imaging During Long-term Culture of Brain Slices in an Enclosed System

To determine how accurately fiducial marks can be utilized to reimage the same cells within the same fields over time, we examined slices grown on photoetched coverslips (Figure 6A). Neurons were visualized by staining with a vital dye (100 nM for 2 h; does not stain non-neuronal cells), which disappears from neurons over time without harming the cells²⁵. We identified a fiducial mark in a single grid square (Figure 6A, B), found a region of vital dye-labeled neurons 24 h after labeling, recorded the x and y offset positions from the fiducial mark (Figure 6B), and collected, using a 60X objective, confocal image stacks of this region, repeating the imaging on 4 consecutive days. The maximum projection images of a 30 µm image stack taken in the same location are shown (Figure 6C-F). Although some morphological changes occur within the region over 4 days, the identical cells (several of which are marked) can be followed over time. The fluorescence intensity of the vital dye declined over time but neurons were still clearly identifiable 4 days after labeling. Although most of the neuronal vital dye fluorescence was diffuse within the cytoplasm, some punctate staining was always observed, which became more noticeable as background fluorescence declined. In unhealthy slices, a punctate staining of non-neuronal cells was also observed as slices deteriorated. These results demonstrate that identified cells in slices can be repetitively imaged by using fiducial marks to find them.

To examine time-dependent changes in neuronal organization and viability within slices during long-term culture, we followed the same slices over 5 weeks, labeling with fluorescent neuronal vital dye once per week 24 h before imaging. Multiple rounds of staining with this vital dye over several weeks increased accumulation of aggregates. Neurons within freshly plated slices that were still in plasma clots were loaded with dye and imaged at one day in vitro (1 DIV). The same slices were imaged again weekly for 5 weeks. Images obtained from a single slice with a 4x objective are shown (Figure 7A-E). The pyramidal cell layers of the CA and DG regions are brightly labeled when excited at 488 nm and fluorescence emission measured at > 620 nm. Over a 5-week period of observing 19 slices, three slices came off the coverslip and two others lost their typical morphology and became opaque, an indication of their death. Thus, a survival rate of about 70% for experiments should be considered, and extra slices prepared to assure an adequate number for analysis. Slices were prepared at a nominal thickness setting on the tissue chopper of 300 µm. After 5 weeks in culture, we measured the slice thickness by imaging neuronal vital dye-stained slices from the coverslip up through the slice with a 40X oil objective on a spinning disc confocal microscope. Loss of focus of neurons occurred at an average of 257 nm (n = 5 slices with multiple locations used per slice), demonstrating that very little thinning of the slice had occurred from the time of plating. We could not accurately measure the slice thickness by fluorescence microscopy at the time of plating because the vital dye entrapped in the plasma clot gave diffuse fluorescence making it difficult to accurately measure the position at which loss of focus occurred. However, 3D images of neurons within slices are easily obtained in slices after the plasma clot is removed. The 21 DIV slice, shown at low magnification in Figure 7F, was imaged with a 60X objective on a confocal microscope (1 µm steps) 3 days after loading with the neuronal vital dye. A 60 µm 3D image was built from the focal planes (Figure 7G). Neurons and their neurite processes that are labeled with the vital dye can be followed in 3D. Morphology and 3D structure of slices were well-maintained over at least 3 months, and the longest times were used in the current study.

Major changes in the morphology of a previously imaged position also occurred in some slices, suggesting that movement of the slice on the coverslip may take place. Certainly, over longer intervals between imaging it became more difficult to know with certainty that the cells in the field being imaged were identical to the ones observed in previous imaging sessions. Thus, maximum projection images of confocal stacks of vital dye-labeled slices acquired at the identical location with the 60x objective at weekly intervals spanning 4 weeks (Figure 8) show that neuronal viability is well maintained but that it is difficult to identify a specific neuron over time when images are obtained with long time intervals between sessions. Presumably the pattern of cells labeled with multiple fluorophores would be more easily recognized, as are cells in localized groups when observed by scrolling through an image stack.

To assess the usefulness of different viral vectors for introduction of exogenous genes in neurons of hippocampal slices, we compared slice infectivity using AV, AAV, and recombinant lentiviral vectors, each expressing different fluorescent tags or using different promoters to drive expression. AV (2 x 10⁷ infectious U/slice) expressing cofilin-mRFP behind a strong, non-cell specific CMV promoter was used to infect a mouse hippocampal slice that had been cultured 9 weeks on coverslips. Expression of cofilin-mRFP was found throughout the slice at 5 days post-infection, with expression most intense around the slice periphery as observed with a 4x objective (Figure 9A). Cells within the slice expressing cofilin-mRFP were also observed with a 20X objective (Figure 9B) with some bright punctate staining and also diffuse expression in both neurons and non-neuronal cells. After 17 weeks in culture (8 weeks post-infection), spontaneous cofilin-rods had formed in some cells, presumably driven by overexpression of wild type cofilin-mRFP (Figure 9C)^26,27.

We also demonstrated that AAV (10¹⁰ particles) could be used for expression in slices. Images of slices infected at 9 weeks in culture with AAV, in which a neuronal specific synapsin promoter drove expression of GCaMP5-cofilin-mRFP with a self-cleaving P2A peptide sequence in the linker of the translated polyprotein²¹, were captured 8 weeks post-infection (17 weeks in culture). In neurons expressing the GCaMP5 and cofilin-mRFP, some cofilin rods/aggregates formed (Figure 9D-F). The fluorescence intensity of the rods/aggregates was so strong that very little fluorescence of a diffuse cofilin-mRFP could be observed without complete saturation and blossoming of the cofilin fluorescence image of rods. Spontaneous cofilin rods appear in neurons in which wild-type cofilin-fluorescent protein chimeras have been over-expressed^26,27, as well as in stressed neurons¹⁰. Based on the titers of adenovirus, which are determined on the basis of infectivity¹⁴ and the particle counts used for determining AAV titer, about 100 to 500 fold higher particle numbers of AAV are needed to obtain approximately the same infectivity/expression in slices compared to AV.

To follow recombinant lentivirus-mediated expression of fluorescent proteins in slices, slices were infected at 6 DIV with 1, 3, 10, and 30 µL aliquots of a recombinant lentivirus for neuronal specific expression (synapsin promoter) of cofilin-R21Q-mRFP, developed as a live cell imaging probe for cofilin-actin rod formation¹⁶. Slices were labeled with neuronal vital dye at 11 DIV and imaged in specific regions for the dye and cofilin-R21Q-mRFP expression on 12 and 14 DIV. The slice infected with the 30 µL aliquot of virus did not survive for imaging but triplicate slices treated with the other volumes of virus showed a dose-dependent expression of mRFP. Figure 10 shows images of slices infected with 1 µL and 10 µL of lentivirus at 6 and 8 days post-infection. Multiple regions of two different slices were quantified for co-staining of neurons with the vital dye and mRFP expression. For slices infected with 1 µL of lentivirus, about 28% of neurons expressed mRFP at 6 days post-infection, increasing to 85% by 8 days post-infection. For slices infected with 10 µL of lentivirus, about 58% of neurons expressed mRFP at 6 days post-infection, increasing to 86% by 8 days post-infection. Thus, 1 µL of the lentivirus was sufficient to provide widespread slice infectivity and neuronal expression by 8 days post-infection.

To demonstrate that this culture system is useful in following development of cofilin pathology, slices infected with lentivirus for expressing cofilin-R21Q-mRFP in neurons were left untreated or treated with various concentrations (1 µM, 333 nM, and 100 nM) of synthetic human Aβ protein that had undergone incubation to form oligomers²⁸. Results from previous studies demonstrated that synthetic Aβ_o induce cofilin-actin rods in up to 25% of dissociated hippocampal neurons^29,30,31. All three slices treated with the 1 µM concentration of Aβ came loose from the coverslip in the first 24 h, whereas all vehicle treated slices (control) and those treated with the 333 nM and 100 nM concentrations of Aβ survived for the two weeks that they were followed. The same cellular regions (CA1, CA3, and DG) in a slice treated with 100 nM Aβ_o were imaged (60x objective) over several days. Control slices that were infected on 6 DIV with lentivirus for expressing synapsin driven cofilinR21Q-mRFP had diffuse cellular mRFP expression by 15 DIV (Figure 11A). Slices exposed to 100 nM Aβ_o at 14 DIV and imaged at 15 DIV showed that the distribution of cofilin-R21Q-mRFP became punctate, appearing in both rod shaped structures and aggregates (Figure 11B). These structures became even more prominent 6 days after Aβ-treatment (Figure 11C, which is the same field of cells as Figure 11B). In many places rich in neurites where neuronal cell somas are absent (Figure 11D), punctate and rod-like arrays of cofilinR21Q-mRFP developed (arrows in Figure 11C), similar to the distribution of cofilin-actin rods previously reported within neurites of Aβ-treated neurons in culture^29,30,31. Thus, this new method for culturing and observing hippocampal slices will allow users to determine the long-term viability of cells in which cofilin aggregates and rods form and the reversibility of the pathology at various stages of development, and to easily perform dose-response measurements on reagents that could block or reverse the formation of cofilin pathology in a more in vivo-like cellular organization.

Figure 1: Preparation of roller tube rack. (A) Template for marking hole positions for drilling out the 15 cm tissue culture dish bottoms. If the figure is printed to the size of the scale bar shown, it can be cut out and used for marking the positions on a 15 cm culture dish for drilling the holes shown. (B) Completed roller tube rack with two tubes inserted. Each rack is numbered on a sticker easily visible on the top of the rack. Please click here to view a larger version of this figure.

Figure 2: Preparation of the roller culture tubes. (A) Front view of the roller tube inserted in a jig on a drill press for drilling out the 6 mm hole in the tube. Dashed line shows the position of the flat sided roller tube in the jig. (B) End view of the jig with the tube inserted and drill bit aligned over hole. (C) Top view of the hole in the jig for drilling out roller tubes with a 6 mm drill bit. White arrow shows the position of the cut-off nail inserted as a stop for positioning the tubes, and black double lines are for bit alignment. (D) Spring clips (black arrow) installed on the jig bottom to securely hold it in position when drilling tubes. (E) After drilling out the hole, the edges are smoothed with a deburring tool and grooves are cut on the inner side of the hole (inset shows the hole viewed through a dissection microscope) to enhance medium draining from the hole during tube rotation. (F) Culture tube with the hole aligned to a hole in the silicone rubber adhesive, to which the coverslip will be attached. Please click here to view a larger version of this figure.

Figure 3: Preparation of hippocampal brain slices. Photos taken with a dissection microscope showing: (A) intact mouse brain. Position of cuts to remove forebrain and cerebellum are shown as blue dashed lines. (B) after removal of forebrain and cerebellum. (C) piece of brain from B is flipped 90° with posterior region (toward the cerebellum) facing up. Positioning the piece next to the side of the dish helps with the removal of the midbrain (blue dashed circle) which can be teased away from the remaining hippocampus, thalamus, and hypothalamus. Two cuts with the forceps (blue dashed lines) allow the remaining piece containing the hippocampus from both hemispheres to be spread flat. (D) The flattened brain piece showing the blood vessel running along the hippocampal fissure (blue arrow). This tissue is placed on plastic film and transferred to the tissue chopper for slicing in the direction of the dashed line. (E) Sliced tissue showing slightly more than half the hippocampus after being returned to GBSS/glucose. (F) Final dissection of the hippocampus and cleaning of the slices to remove non-hippocampal material. (G) Several floating slices after final clean-up. (H) Enlarged photo of a single slice for transfer to coverslip. Please click here to view a larger version of this figure.

Figure 4: Plating and incubating slices. (A) A mouse hippocampal slice is removed from the culture dish on the tip of a spatula using the tip of forceps to help lift it free from the solution. (B) The slice is placed flat in the center of a photoetched and treated 12 mm coverslip on 2 µL of chicken plasma and another 2.5 µL of a 1:1 plasma/thrombin mixture is added to generate a clot. (C) After the clot is set (about 1-2 min), the covering is removed from the silicone rubber adhesive circle on a roller tube and the coverslip is positioned with the clot centered into the hole; then the coverslip is pressed in place with a thumb and held in position for about 1 min. (D) Add 0.8 mL of complete culture medium. (E) Roller tube holders inside of a large roller incubator with the front raised to tilt 5° to keep the medium at the bottom of the tubes. Please click here to view a larger version of this figure.

Figure 5: Roller tube holder and stage plate incubator. (A) Tube holder that positions the tube such that the coverslip is maintained in position for imaging. (B) Tube holder mounted in a microscope stage adapter plate. Sliders on the side hold the tube securely for imaging. (C) Stage adapter with the tube holder and side panels added containing heating strips connected to a thermoelectric temperature controller. Once tubes are mounted and positioned, a solid top to the box can be added to help maintain the temperature during imaging. Orange wire is the thermocouple lead. Plans for the design and building of the stage adapter and heater are available at: https://vpr.colostate.edu/min/custom-machining/roller-tube-holder/.  Please click here to view a larger version of this figure.

Figure 6: Using fiducial marks for repetitive imaging of same cells. (A) Hippocampal slice on photoetched coverslip (1 mm squares) with subiculum unfolded (tail) viewed with 4x objective and bright field illumination. Curvature of plastic roller tube helps create an oblique illumination that enhances the visualization of the grid. The box shows the position and size of a 60X field. (B) A view of the same slice with a 20x objective for finding the fiducial mark as the tip of the bottom of the 4 from square 34. The y and x offsets are shown to reproducibly locate the center of the desired box for higher magnification confocal imaging. (C–F) A slice labeled with neuronal vital dye 13 DIV was imaged using a 60x objective and making a 30 µm projection image on 4 consecutive days (14-17 DIV). Identical neurons were imaged each day. The position of the nucleus in each of three neurons is marked with a different symbol. They are more easily identified by scrolling through image stacks as their 3D position changes slightly. Please click here to view a larger version of this figure.

Figure 7: Imaging neurons in slices with a neuronal vital dye. (A) Neuronal vital dye stained slice in plasma clot taken 24 h after plating with 4X objective. Dye (100 nM) was added when the slice was first placed in the roller tube holder and washed out 2 h later. Subiculum is curled around the hippocampus in the clot. (B–E) Following clot dissolution by plasmin added at 6 DIV, the slice was reloaded 5 times with the vital dye 24 h in advance of imaging at weekly intervals. Images shown were collected at 8, 21, 28, and 35 DIV (B–E, respectively). (F) Hippocampal slice cultured for 3 weeks and stained with neuronal vital dye 24 h in advance of imaging with 4X objective. (G) Confocal stack of images on the same slice as in F showing a 3D view of 61 planes taken at 1 µm intervals. Neuronal vital dye clearly labels both neurites and cell body but it is excluded from the nucleus. Please click here to view a larger version of this figure.

Figure 8: Repetitive imaging of neurons in a single location of the slice over 4 weeks. Maximum projection images of 30 µm confocal image stacks of the same field of cells (by positioning) from vital dye-labeled slices taken at 12, 20, 30, and 40 DIV (A–D, respectively). It is difficult to repetitively identify individual cells over the longer time frames in projection images. However, even over these long periods, identification of the same cells is often possible by scrolling through the image stacks or building 3D images that can be rotated, such as shown in Figure 7G. Please click here to view a larger version of this figure.

Figure 9: Viral-mediated expression and imaging of fluorescent proteins. (A) Hippocampal slice cultured for 9 weeks and infected with adenovirus for expressing cofilin-mRFP behind a CMV promoter. Expression was found throughout the slice at 5 days post-infection but fluorescence was brightest near the slice periphery. (B) Same slice showed expression in cells deeper within the slice when viewed with 20X objective. Image is a projection from a stack of 20 images spaced 2 µm apart. (C) The same slice was examined after 17 weeks in culture (8 weeks post-infection) and cofilin mRFP was observed in rod shaped aggregates as seen in this projection image from a 70 µm stack of 23 images, 3 µm apart, taken with a 40X objective. (D–F) Mouse hippocampal slice infected at 9 weeks in culture with an AAV expressing a GCaMP5-(P2A)-cofilin-mRFP behind a synapsin promoter. Fluorescence was visible in both red and green channels after 10 days. A single plane image of the slice showing the expression of (D) GCaMP5, a calcium sensitive reporter, (E) many cofilin-containing rods, and (F) an overlay image. Please click here to view a larger version of this figure.

Figure 10: Expression of cofilin-R21Q-mRFP driven by a synapsin promoter in neurons infected with recombinant lentiviral vectors. Detection of a weak fluorescence signal is first observed by about 3-4 days post-infection using 10 µL of virus and becomes usable by 5-6 days, (E) as seen in these images acquired with a 60x objective. Although it takes longer to achieve the same levels of expression with 1 µL of virus, by 8 days post-infection a similar high percentage of neurons (vital dye labeled) were expressing the cofilin-R21Q-mRFP. Only 27% of neurons were positive for mRFP fluorescence at 6 days post-infection with 1 µL (A, B) but this increased to 85% (C, D) by 8 days. Please click here to view a larger version of this figure.

Figure 11: Aβ oligomer-induced cofilin pathology in mouse hippocampal slices. All images taken as 30 µm image stacks with a 60x objective and are shown as maximum projection images. Slices were infected with cofilin-R21Q-mRFP at 6 DIV. (A) 15 DIV slice treated on 14 DIV with vehicle (DMSO/HAMS F12 medium used to generate Aβ_o). (B) Slice 15 DIV treated with 100 nM Aβ_o. (C) Same field as in B taken at 20 DIV and shown in (D) as an overlay with neuronal vital dye label. Arrows show linear arrays of cofilin aggregates and rods in the region of the slice containing neurites but few cell bodies. Please click here to view a larger version of this figure.