Fragmentation as a Mechanism for Growth Cone Pruning and Degeneration

Abstract

During early development of the central nervous system, there is an excessive outgrowth of neuronal projections, which later need to be refined to achieve precise connectivity. Axon pruning and degeneration are strategies used to remove exuberant neurites and connections in the immature nervous system to ensure the proper formation of functional circuitry. To observe morphological changes and physical mechanisms underlying this process, early differentiating embryonic stem cell-derived neurons were used combining video imaging of live growth cones (GCs) with confocal laser scanning microscopy and atomic force microscopy, both on fixed and living neurons. Using this method, we could highlight the presence of submicrometric fragments in still and in some of the retracting GCs. The observed fragmentation is not an artifact of atomic force microscopy scanning or fixation, or the result of apoptosis. Therefore, the morphology of GCs depends on their overall motility, and fragmentation seems to be the fate of GCs that have not found a correct destination.

Introduction

T he formation of the nervous system has fascinated researchers for over a century. Understanding the formation of this ordered complexity is one of the greatest challenges of modern science. The molecular mechanisms by which progenitor cells differentiate and acquire their cell identity are just beginning to be fully understood [1,2]. During the early development of the central nervous system, there is an excessive outgrowth of projections, which need to be refined to achieve precise connectivity [3]. The selective elimination of neuronal cell processes, or neurites, is an essential step during normal development and under certain pathological circumstances [4]. Growth cones (GCs) located at the tip of developing and regenerating neurites are responsible for neurite extension, axonal path finding, and target cell selection. A growth-permissive environment facilitates GC assembly, whereas various inhibitory molecules induce GC collapse and neurite retraction [5].

Moreover, exuberant and/or erroneous neuronal connections need to be pruned to achieve precise connectivity. Pruning occurs through retraction, degeneration, or a combination of both [4].

GCs express receptors for extracellular guidance cues and integrate this information into directional movement toward the target cell [6], thus acting as sensors, signal transducers, and motility devices. Filamentous (F)-actin is the primary cytoskeletal element that maintains the GC shape and is essential for proper axon guidance, whereas microtubules are essential for giving the axon structure and serve an important function in axon elongation [7]. Although the role of guidance molecules [8] for GC movements, underlying signaling of both the F-actin and microtubules, is well established [7,9,10], the fate of GCs that have not reached a correct final destination is still not completely known. Biochemical mechanisms that control pruning in invertebrates and in the peripheral nervous system of vertebrates are beginning to be understood, but those in the central nervous system remain more elusive [3]. However, relatively little is known about morphological changes and physical mechanisms underlying this process.

In vitro differentiation of embryonic stem (ES) cells recapitulates early events in the development of the mammalian nervous system: ES cells can both generate and respond in vitro to signals that normally regulate embryonic development [11,12]. Moreover, functional networks can be obtained with ES cell-derived neurons [13]. Therefore, ES cells provide a useful model to study in vitro early mammalian development.

In the present article, we have combined video imaging of GC motion with confocal laser scanning microscopy (CLSM) and atomic force microscopy (AFM) to investigate the topography of early differentiating ES cell-derived neurons. GCs actively exploring the environment before fixation had a smooth external surface. In contrast, GCs that were immobile before fixation revealed a fragmented shape: the overall shape of the GC was still recognizable but instead of being compact it was composed of several nanoscale structures either partly attached or completely isolated from the rest of the GC. In addition, using AFM on live specimen, we noticed fragmentation in some of the retracting GCs. Therefore, the morphology of GCs depends on their overall motility, and fragmentation seems to be the fate of GCs that have not found their appropriate target.

Materials and Methods

Coverslips and printing

Coverslips with printed markers were fabricated starting from 24-mm-diameter glass coverslips (Menzel-Gläser) by optical lithography and metal evaporation. A polymeric photoresist was deposited on one side of the coverslips by spin coating. The indexed pattern was produced by UV exposure through a patterned optical mask. A 20-nm titanium layer was finally deposited and stripped by lift-off techniques. The glass coverslips were cleaned in acetone and methanol and then autoclaved to remove possible contamination produced by the lithographic process. To avoid possible interference of the printed grid on cell growth, cells were plated on the unprinted side of the glass coverslip.

ES cell culture

ES cells were induced to differentiate into neurons as previously described [13]. Undifferentiated mouse BF1/lacZ ES cells were plated on Mitomycin C (Sigma-Aldrich)-treated MS5 cells as single-cell suspension at a density of 250 cells/cm² in knockout serum replacement (KSR; Invitrogen) and cultured for 6 days. The KSR medium was changed every day. The ES cell-derived epithelia structures were then mechanically separated from MS5 cell monolayer by incubating the cells with Hank's balanced salt solution (Sigma-Aldrich) supplemented with 0.15 M HEPES (Sigma-Aldrich) for 20 min and flushing through the pipette tip near the colony without removing the MS5 feeder. The detached ES cell colonies were resuspended with KSR medium and plated on 15 μg/mL polyornithine (Sigma-Aldrich) + 1 μg/mL fibronectin (Invitrogen)–coated dishes. After 2–3 h, the medium was changed with N2 medium (Invitrogen) containing 10 ng/mL basic fibroblast growth factor (bFGF; R&D Systems) and 1 μg/mL fibronectin (amplification medium). Cells were induced to proliferate in the presence of bFGF for 4 days, and during the last 2 days, 200 ng/mL of Sonic Hedgehog (R&D Systems) and 100 ng/mL FGF 8 (R&D Systems) were added as patterning factors. Cells were then trypsinized, resuspended in 10% FBS N2 medium to neutralize trypsin, and then plated at a density of 3 × 10⁴ cells/cm² on 15 μg/mL polyornithine + 1 μg/mL laminin–coated 24-mm-diameter glass coverslips with a printed grid in N2 medium to induce differentiation. Twenty-four hours after differentiation induction, cells were fixed in 4% paraformaldehyde (Sigma-Aldrich) containing 0.15% picric acid (Sigma-Aldrich) in phosphate-buffered saline (PBS; Invitrogen). In the experiments in which different fixation methods were tested, the following conditions were used: (1) 4% paraformaldehyde–0.25% sucrose for 20 min at room temperature; (2) 0.25% glutaraldehyde in 100 mM cacodylate (pH 7.4)–5 mM CaCl₂–10 mM MgCl₂ at 37°C for 15 min.

Video imaging

Twenty-four hours after differentiation induction, ES cell-derived neurons were analyzed with time-lapse differential interference contrast (DIC) microscopy. Coverslips were placed in the observation chamber inside the incubator box combined with a precision air heater that ensures tightly controlled temperature of both specimen and microscope and additional CO₂ control (Life Imaging Services). Cells were kept at 37°C and 5% CO₂ in N2 medium. DIC images of moving ES cell-derived GCs were obtained with a Leica DMIRE2 confocal microscope (Leica Microsystems GmbH) equipped with DIC and fluorescence optics, diode laser 405 nm, and Ar/ArKr 488 nm and He/Ne 543/594 nm lasers. Samples were imaged with a 40 × magnification and a 1.25 NA oil-immersion objective at a single focal plane. Images with resolution between 512 × 512 and 1,024 × 1,024 pixels were acquired every 10 s, for a total of 80 frames (total duration 13 min and 20 s), and the motion was recovered using an operator-assisted program. Immediately after video imaging, cells were fixed with 4% paraformaldehyde containing 0.15% picric acid in PBS for 20 min at room temperature (20°C–22°C) and then kept at +4°C. The grid on glass coverslips was the reference used to return to the selected locations after AFM analysis.

Confocal imaging

Cells were fixed in 4% paraformaldehyde containing 0.15% picric acid in PBS, saturated with 0.1 M glycine (Sigma-Aldrich), permeabilized with 0.1% Triton X-100 (Sigma-Aldrich), saturated with 0.5% bovine serum albumin (Sigma-Aldrich) in PBS, and then incubated for 1 h at room temperature (20°C–22°C) with primary antibodies: rabbit polyclonal antibody against neural cell adhesion molecule (NCAM; Sigma-Aldrich) and mouse monoclonal antibody against neuronal class III β-tubulin-TUJ1 (Covance). The secondary antibodies were goat anti-mouse-biotinylated (Dako) and anti-rabbit-594 Alexa (Invitrogen). F-actin was marked with Alexa Fluor 488 phalloidin, whereas biotin was recognized by Marina Blue-Streptavidin (Invitrogen) and incubated for 30 min at room temperature (20°C–22°C). Total nuclei were stained with 2 μg/mL in PBS Hoechst 33342 (Sigma-Aldrich) for 5 min at room temperature. The cells were examined using the same confocal instrument described in the previous section. The fluorescence images (1,024 × 1,024 pixels) were collected with a 63 × magnification and 1.4 NA oil-immersion objective. Leica LCS Lite and Image J by W. Rasband (developed at the U.S. National Institutes of Health and available at http://rsbweb.nih.gov/ij/) were used for image processing. Registration of confocal images was obtained as previously described [14].

AFM imaging

AFM was performed using the Nanowizard II instrument (JPK) combined with an inverted optical microscope (Zeiss Axiovert 200) and a fluorescence setup (Zeiss X-cite).

For imaging fixed cells, AFM was operated in contact mode in liquid, adjusting the contact force during imaging to compensate the thermal and chemical bending of the cantilever and to minimize the force exerted by the tip on the sample during scanning. Gold-coated, soft V-shaped silicon nitrite cantilever (VEECO Metrology) with nominal force constant of 0.01 N/m, a resonance frequency in the range of 4–10 kHz, and length between 305 and 315 μm were used. The tip, 2.5–8 μm in height, had a nominal tip radius of curvature of around 20 nm. Forces were kept between 100 pN and 1 nN during scanning. Coverslips were mounted in the AFM liquid cell, immersed in PBS containing 2 μg/mL gentamycin (Sigma-Aldrich) to avoid bacterial contamination, and positioned on the xy stage. A 0.6 NA objective and 40 × magnification were used to find specific GCs using the position reference markers. After laser alignment and tip calibration, and before starting the process of image acquisition, the system was left to settle for half an hour with the laser and the microscope condenser turned on, until thermal and force drift reached a steady state. The scan speed of the tip was maintained below 5 μm/s to minimize the tip-induced damage. AFM images were acquired with a resolution between 512 × 512 and 1,024 × 1,024 pixels. Postprocessing was performed using JPK Image Processing or, to obtain detailed data analysis, using a Scanning Probe Imaging Processor Image postprocessing (Image Metrology A/S). Confocal and AFM images were superimposed following the registration procedure previously described [14]. For imaging living cells, AFM was operated in tapping mode at 0.6 scan line per second and with a maximum of 256 scan lines, using soft cantilevers and a low feedback gain. Gold-coated, silicon nitride cantilevers (VEECO Metrology) were used with nominal spring constant between 0.03 and 0.009 N/m and length between 50 and 70 μm. The tip, 5–10 μm in height, was square pyramidal in shape, with a nominal tip radius of curvature of around 25 nm. A resonance frequency in the range of 35–40 kHz, scan speed of 6 μm/s, and free oscillation amplitude of 500 nm were determined in the proximity of the cell surface. Coverslips with living cells were mounted onto a stage (Biocell™; JPK) in differentiation medium, maintaining the temperature at 37°C. In these conditions, ES cell-derived neurons survived up to 4–5 h.

Results

We used CLSM combined with immunofluorescence to stain and map the distribution of different cellular components in GCs. However, because of their small dimensions, conventional optical instruments do not allow to characterize their morphology and visualize structural details smaller than 200 nm. Therefore, we combined CLSM with AFM [15] to obtain a precise measurement with a nanometer resolution of the shape and size of the cellular surface.

AFM scanning reveals different GC morphologies

ES cell-derived neuronal precursors were obtained using the protocol previously described [13]. Cells were plated on coverslips and induced to differentiate for 24 h. During this period of culture, ES cell-derived neurons extended neurites with GCs moving forward, retracting, and exploring the environment with their filopodia. The structure of the differentiating GCs was analyzed by AFM in contact mode in liquid. Cells were plated at a density of 3 × 10⁴ cells/cm² to obtain isolated GCs and avoid overlapping structures. This density, however, was sufficient for neuronal survival in culture.

At this stage of differentiation, very different structures of GCs were observed (Fig. 1). The diameter of these GCs varied from 1.5 to 28 μm and their height varied from 65 to 593 nm. Some GCs (Fig. 1A–C) appeared swollen and smooth, with several filopodia spreading from the central domain. The height profile differed substantially from 1 GC to another, but in 81 of 119 GCs analyzed, it always exceeded 200 nm (Fig. 1D). Twenty-five of 119 GCs were flat with few or no filopodia (Fig. 1E–G). Their height profile was consistently below 200 nm (Fig. 1H). Thirteen of 119 GCs showed a ruffled and fragmented structure with several holes (Fig. 1I–K). The height profile of these GCs was also below 200 nm, and their thickness almost vanished in some regions (Fig. 1L).

FIG. 1.

High-resolution AFM height images of embryonic stem cell-derived GCs fixed after 24 h of differentiation. GCs could have swollen and smooth (A–C), flat and rough (E–G), or fragmented (I–K) structures. Scale bar: 2 μm. Color bar: from 0 to 800 nm (A–C, E–G) and from 0 to 200 nm (I–K). (D, H, L) Height profiles obtained along the white lines for the images in A–C, E–G, and I–K, respectively. AFM, atomic force microscopy.

Holes and fragments of GCs are not artifacts of AFM scanning or fixation

During image acquisition, the AFM tip interacts with the cell surface and could modify or damage the cell membrane. To rule out the possibility that the fragmented structures were the result of damage induced by the AFM tip, a series of images of a compact GC at increasing tapping forces of 200, 1,000, and 3,000 pN, respectively, were acquired (Fig. 2A–C). The shape of the GC was not modified at increasing scanning forces and no tip-induced fragmentation of the membrane was observed. The height profile analysis of the GCs at varying forces showed a decrease in height (Fig. 2F–H) caused by a compression of the cell structures, probably due to the residual cell elasticity, after the fixation procedure. However, when the imaging force was restored to the initial value of 200 pN, the cell morphology was exactly the same as in the first scan (see the nearly perfect superimposition of the grey shadowed and dotted curves in Fig. 2F–H). Although the debris on the right of the GC was removed by iterated scans, the GC shape appears unchanged (compare panels A and D). Figure 2 shows the results of a representative experiment among the 4 performed.

FIG. 2.

AFM error images of the same GC obtained with increasing scanning forces: (A) 200 pN, (B) 1,000 pN, (C) 3,000 pN, and (D) again 200 pN. (E) AFM topography image acquired in D. Scale bar: 4 μm. (F–H) Height profiles along white lines in E for all analyzed forces. Grey shadowed profiles represent the initial height at 200 pN scanning force; black and grey continuous lines were acquired at 1,000 and 3,000 pN, respectively. Dotted lines were obtained when returning to 200 pN.

Another possible source of artifact could be the paraformaldehyde fixation. Therefore, we tested other 2 fixation methods and looked for the presence of fragmented GCs. Fragmented structures were found with all the fixation methods that were used (Fig. 3). Both fragmented and compact GCs were present on the same coverslip and sometimes on neighboring cells, regardless of the fixation protocols used, giving an indication that fragmentation was independent from fixation.

FIG. 3.

Topography and height profile of embryonic stem cell-derived GCs fixed with (A) 0.25% glutaraldehyde, (B) 4% paraformaldehyde and 0.15% picric acid, or (C) 4% paraformaldehyde and 0.25% sucrose (see Materials and Methods section) at 24 h of differentiation. Scale bar: 3 μm. (D–F) Height profiles along the white line shown in the above AFM images confirmed the presence of holes (height value corresponding to 0 nm) for all fixation methods used.

The ultimate proof that these structures are not artifacts came from finding holes and fragments in GCs of living neurons analyzed by AFM (Fig. 4). The structure of live GCs was analyzed by AFM in tapping mode. The acquisition of AFM images of GCs required several minutes, and because of the movement of GCs and the fluidity of the membrane, a high-resolution AFM image cannot be achieved. Under these conditions, however, it was possible to follow the GCs exploring their environment and changing their morphology in time. In this way, we were able to follow 23 living GCs, obtaining several consecutive AFM images. Some (14 of 23 imaged cells) GCs preserved compact and smooth surfaces during the scanning period (Fig. 4A) and cycles of protrusion and retraction were observed, indicating that the physiological dynamics was poorly affected by AFM scans and that the cantilever tip did not damage the GCs (Fig. 4A, D). In 6 cases, the GCs retracted leaving behind fragments with variable dimensions, and in 3 cases we observed the presence of holes in the GCs (Fig. 4B, C). In one case the holes were already present at the first scan.

FIG. 4.

AFM topography and height profile of live GCs. Scale bar: 3 μm. (A) Compact and (B, C) fragmented GCs were observed. (D–F) Height profiles along a white line in the above panels show the holes (height value corresponding to 0 nm) observed in fragmented GCs (B, C).

Altogether, these observations suggest that the fragments observed on the fixed cells were originated before fixation, and therefore, fragmentation is a physiological phenomenon.

Fragmentation is a characteristic of immobile GCs

We hypothesized that the different morphologies observed in the GCs might correlate with different types of movement. To verify this idea, we combined time-lapse microscopy on living cells with AFM scanning and immunofluorescence after fixation. To retrieve the same cell for the different assays, coverslips with printed markers were prepared, on which precursors were plated and induced to differentiate. The marker grid on the coverslip allowed retrieving the same GC after immunofluorescence assay or when analyzed with different microscopes. The motion of >100 GCs was followed by video imaging with time-lapse DIC imaging for about 10 min. From the time-lapse analysis, we could categorize 4 different classes of movements: exploring (43/104), growing (16/104), retracting (17/104), and stasis (28/104); 29 of these were also imaged with AFM, followed by staining with appropriate antibodies.

Exploring (9/29)

GCs of ES cell-derived neurons could explore very efficiently the surrounding free space (Fig. 5A, B and Supplementary Video S1; Supplementary Data are available online at www.liebertonline.com/scd) and the tip of their filopodia could move with a velocity up to 80 nm/s, with an average value of 25 ± 7 nm/s. As shown by the DIC images taken at 0 and 11.5 min time intervals (Fig. 5A, B), the majority of filopodia moved significantly from frame to frame (10-s interval) and independently in all directions. Some filopodia were continuously extending and retracting, changing rapidly in their length, as indicated by the arrow in Fig. 5A, B, whereas some others were freely exploring, maintaining their shape. These GCs had a smooth and compact surface with F-actin mainly localized in the peripheral domain and filopodia, whereas β-tubulin III stained the neurite, the central domain, and the transition zone (Fig. 5E). The GCs had an average height of 852 ± 235 nm. Filopodia branching from these GCs had a height varying between 67 and 219 nm (n = 40; average height: 98.8 ± 48.3 nm).

FIG. 5.

Exploring and growing GCs. (A, B) DIC images showing filopodia movement (see arrow) at 0 and 11.5 min, respectively. (C, D) DIC images showing growth progression (see arrow) at 0 and 6 min, respectively. (E) Registered confocal image texture stained for F-actin (green) and β-tubulin III (blue) and mapped on 3D AFM image for the exploring GC shown in A. (F) Registered confocal image texture stained for F-actin (green) and NCAM (red) and mapped on 3D AFM image for the growing GC shown in C. The superimposed AFM and confocal images were obtained following the registration procedure previously described [14]. Scale bar: 10 μm. DIC, differential interference contrast. NCAM, neural cell adhesion molecule.

Growing (5/29)

Neurites could grow by 1–5 μm in 2–10 min and in 2 cases a neurite grew up to 14 μm in <2 min (Fig. 5C, D and Supplementary Video S2). After fixation, AFM revealed a compact and smooth surface (Fig. 5F) in all growing GCs (n = 5). Filopodia height reached almost 2 μm and their average length was 3.5 μm (n = 20). Immunofluorescence assay confirmed the expression of neuronal tubulin, F-actin, and NCAM.

Retracting (7/29)

In 7 neurons analyzed by time-lapse, GCs retracted their filopodia and the whole neurites from the original position (Fig. 6A, B). Retracting GCs moved with a velocity ranging from 80 to 135 nm/s. Some retracted filopodia had a pearl-like morphology where their diameter and height increased and decreased along their length in a pearl necklace fashion (Fig. 6C). GCs expressed β-tubulin III (not shown) and NCAM (Fig. 6D). After fixation, AFM scanning revealed the presence of fragments near the tip of fast-retracting GCs. To confirm whether these fragments were actual parts of the GCs left behind by the neuron, AFM of living GCs was performed in tapping mode (see Materials and Methods section). An example of a retracting GC observed for 160 min is shown in Fig. 6E. During retraction, the GC height varied between 100 and 200 nm, progressively thinning (see height profiles measured at the level of the white scan-line). After 40 min the region of the GC indicated by the white arrowhead retracted, forming an invagination. The 2 filaments, possibly cytoskeletal components, which were initially detected (see white arrows), completely disappeared after 40 min. At 160 min, a fracture appeared in the neurite and the neurite completely retracted, leaving behind a large isolated fragment (see grey arrow). In all 6 retracting GCs observed with live AFM imaging, fragments with an average diameter of 85 ± 27 nm and an average height of 75 ± 28 nm were detected.

FIG. 6.

Retracting GCs. (A, B) DIC images showing the retraction of the filopodia (see white arrows) at 0 and 7.5 min, respectively. (C) AFM topography and (D) registered [14] confocal image of the same GC stained for NCAM. (E) AFM topography images of a fragmented GC acquired under live condition at time points indicated on each panel: 0, 20, 40, 80, 120, and 160 min, respectively. The contour of the GC at time 0 min (white line) is superimposed to the GC images recorded at increasing time. White arrowheads indicate invagination. White arrows indicate filaments initially observed and then disappeared in following images. The grey arrow indicates the fracture between GC and neurite. Height profiles along the white line (shown only at time 0) are plotted for each time point. Scale bar: 5 μm.

Static (8/29)

GCs were classified as static when the external contour of the GC remained in the same position and the length of their neurite did not change during observation. Some activity of exploring filopodia in the region close to the neurite was observed, indicating that the neuron was alive despite the lack of growth/retraction of the neurite. Immobile GCs (8 cells analyzed) had an apparently intact shape when viewed with time-lapsed DIC images (Fig. 7A), but their 3D shape was highly fragmented when viewed with AFM (Fig. 7B). These fragments contained actin filaments but not tubulin, and they were positive for NCAM (Fig. 7C), indicating that fragments left behind by GCs are formed by chunks of actin filaments enveloped by the cellular membrane (see registered confocal image in Fig. 7C, F).

FIG. 7.

Static GCs. (A) DIC, (B) AFM topography, and (C) registered [14] confocal image of the same GC stained for F-actin (green), NCAM (red), and β-tubulin III (blue). Scale bar: 5 μm. (D) Height distribution of the fragments from all observed fragmented GCs. Height measured at the median point. (E) Height distribution of the filopodia from all observed GCs. Height measured at their medial length. (F) Confocal image of the inset shown in C. (G) Height of the membrane surrounding holes versus hole area for 4 fragmented GCs analyzed. (H) Fragment height versus fragment average radius for 4 fragmented GCs analyzed. (I) Height profile along the white line shown in F.

Fragments contain actin filaments

Profiles of AFM images through these fragments (Fig. 7I) showed that their height varied from 50 to 150 nm and that they were isolated from each other. Height of fragments and filopodia had a similar distribution varying from <30 up to 300 nm (Fig. 7D, E). Immobile GCs not only were surrounded by fragments, but they also had holes (see Materials and Methods section). GC regions surrounding these holes had a height varying from 20 to 90 nm and holes had an area varying between 0.03 and 0.650 μm². Fragments had a height varying from 40 to 400 nm, with an area varying between 0.4 and 6 μm². The height of the membrane surrounding holes is weakly correlated to their area (Fig. 7G), whereas the height of small fragments increases linearly with their radius (Fig. 7H).

The observed fragmentation was not associated with apoptosis because none of the neurons with static GCs showed signs of membrane blebbing or cell shrinking typical of apoptotic cell death during the time-lapse DIC imaging. Moreover, none of the eight cells that were also analyzed by AFM and immunofluorescence after fixation had an apoptotic nucleus as observed by Hoechst staining (Supplementary Fig. S1). We can therefore conclude that GC fragmentation is not a part of the apoptotic process.

Fragmentation was a characteristic of static GCs. It could be that the cells stop advancing in the absence of positive stimuli from other cells and afterward collapse, but the same mechanism could occur during pruning when an erroneous contact need to be removed. To verify this hypothesis, we performed time-lapse confocal microscopy on living neurons that made contact (n = 21) and verified that their GCs did not move during the observation time. AFM images of the region of contact revealed compact structures with a smooth surface in 95% of cases (Fig. 8B), including the region where the 2 filopodia made contact (Fig. 8D). They were positive for both β-tubulin III and F-actin, as confirmed by the immunofluorescence assay (Fig. 8C). However, in 5% of the cases we observed fragmented structures (1/21 cells), indicating a similar mechanism for degeneration of GCs lacking positive stimuli from the environment and for pruning of erroneous or inefficient contacts.

FIG. 8.

GCs that established contact. (A) DIC image of the 2 GCs. (B) AFM height image of the region corresponding to the square in A. (C) F-actin (green) and β-tubulin III (blue) staining of the same cell. Confocal image has been registered onto the AFM image in B. Scale bar: 10 μm. (D) Height profiles of 3 scan lines shown in B.

Discussion

In the present article, we implemented the combination of AFM and fluorescence confocal microscopy described in our previous work [14] by adding time-lapse DIC and AFM imaging under live conditions. Using these 3 different imaging techniques separately, at different times, on different instruments—but on the same samples—we provide a morphological characterization of ES cell-derived GCs related to their movement prior fixation. Using this method, we could highlight the presence of fragments in still and in some of the retracting GCs.

The fragmented structures were not artifacts of AFM scanning or fixation. Although early investigations using AFM [16,17] have shown that the cantilever tip can damage neurons and has been also used to perform nanosurgery, we could demonstrate by increasing measuring forces from 200 to 3,000 pN that the fragmentation of GCs observed is not the result of scanning (Fig. 2); GC fragmentation cannot be attributed to the fixation procedure (Fig. 3) and GC fragmentation is observed also under living conditions (Fig. 4) with comparable proportion (11% fragmented GCs in fixed and 13% in living conditions).

Holes with a micrometric size were previously observed in in vivo studies of GCs from retinal ganglion cell axons [18,19]. These “holes” appear in the spread GCs and are predicted to form from the fusion or contact of lamellar extensions of the GC as they enfold radial glial processes. However, although morphologically similar, these 2 findings reflect a different biological phenomenon: first, because our observations were the result of a technical approach measuring at nanometric level and, second, because we never found fragmented structures in growing GCs.

GC fragmentation was not the result of apoptosis. In fact, none of the 104 cells observed with time-lapse DIC imaging showed signs of cell death, including the 28 static GCs. Moreover, in all the observed cases of fragmented GCs, the nuclei were not apoptotic. This phenomenon resembles the GC collapse induced by several factors such as mercury [5], X-ray [20], or semaphorin in the absence of growth factors [21]. The fragmentation of the GCs observed is a local phenomenon, similar to that observed for lysophosphatidic acid induction of collapse in vitro, which, in contrast to other collapsing treatments, is reversible and not toxic [22]. However, to our knowledge, this is the first time that this collapse-like phenomenon was observed to occur spontaneously.

Previous investigations have shown the formation of migration tracks resulting from the release of cellular material onto glass surfaces and artificial matrices for a number of cell types including fibroblasts of different organisms [23,24], mammalian thymoma and sarcoma cells, cytotoxic T-lymphocytes, primary chondrocytes [25], and keratinocytes [26]. Macroaggregates left behind migrating keratinocytes contain high amounts of β1 integrin and parts of the fibronectin and laminin receptors. However, they lack cytosolic proteins including actin and the adhesion complex constituents talin and vinculin. In our experimental conditions, fragments left behind by GCs are composed of cell membrane and also cytoplasmic proteins such as F-actin, as it was reported for other migrating cells [23]. Fragments could correspond to filopodia originally present on the GCs in that position also because the height distribution of filopodia and fragments was similar (Fig. 7D, E), suggesting that they may be composed of similar building/dismantling blocks. The existence of building blocks could agree with a previous study [15] in which it was found that the height of a hippocampal GC corresponds to a multiple of the heights of individual filopodia, possibly because of overlying actin bundles arising from different filopodia.

The release of a fragment might be energetically advantageous for faster retraction and/or change in growing direction compared with recycling of distal elements. However, it cannot be excluded that fragments might act as guidance signals for neighboring neurons.

Vertebrate semaphorins are either secreted or associated with the cell surface. Therefore, they can mediate both long- and short-range (or contact-mediated) signals. Some migrating cells and axons express both receptors and ligands on the cell surface or secrete semaphorins in an autocrine fashion. In vitro and in vivo experiments have implicated semaphorins in the guidance of elongating axons and dendrites, as well as in axon branching, axon pruning, and axon degeneration [21]. The primary role of Sema 3A in the nervous system is to repel GCs from inappropriate areas and to help steer both axons and migrating cells along the correct trajectory [27]. When added in bath, negative guidance molecules cause rapid collapse of GCs characterized by a loss of lamellipodia and filopodia, followed by axon retraction [28].

The fragmentation of GCs that have not established contact or by pruned contact here observed could therefore serve as a migrating track for other neurons, by exposing semaphorins or other membrane proteins that act as receptors and/or ligands for axon guidance.

Footnotes

Acknowledgments

The authors thank Shripad Kondra and Jummi Laishram for their help with image registration and Alpan Bek for useful suggestions about AFM. This work was supported by the EU projects: NEURO Contract No. 012788 (FP6-STREP, NEST) and NanoScale Contract No. 214566 (FP7-NMP-2007-SMALL-1). The authors acknowledge the financial support via an FIRB grant D.M.31/03/05 from the Italian Government, Contr. RICN No. 011936 BINASP from the European Community, funds from the Istituto Italiano di Tecnologia (Research Unit IIT), and the GRAND grant from CIPE/FVG by the Friuli Venezia Giulia region.

Author Disclosure Statement

The authors declare that they have no competing financial interests.

References

Gilbert

. 2000. Neural crest cells and axonal specificity. Developmental Biology. Gilbert

. Sinauer: Sunderland, MA, 407–441.

Lai

, Johnson

. 2008. Neurogenesis or neuronal specification: phosphorylation strikes again! Neuron, 58:3–5.

Faulkner

, Low

, Cheng

. 2007. Axon pruning in the developing vertebrate hippocampus. Dev Neurosci, 29:6–13.

Franze

, Gerdelmann

, Weick

, Betz

, Pawlizak

, Lakadamyali

, Bayer

, Rillich

, Gögler

, Lu

, Reichenbach

, Janmey

, Käs

. 2009. Neurite branch retraction is caused by a threshold-dependent mechanical impact. Biophys J, 97:1883–1890.

Leong

, Syed

, Lorscheider

. 2001. Retrograde degeneration of neurite membrane structural integrity of nerve growth cones following in vitro exposure to mercury. Neuroreport, 12:733–737.

Grzywa

, Lee

, Suter

. 2006. High-resolution analysis of neuronal growth cone morphology by comparative atomic force and optical microscopy. J Neurobiol, 66:1529–1543.

Dent

, Gertler

. 2003. Cytoskeletal dynamics and transport in growth cone motility and axon guidance. Neuron, 40:209–227.

Dickson

. 2002. Molecular mechanisms of axon guidance. Science, 298:1959–1964Erratum in: Science (2003) 299:515.

Allen

, Chilton

. 2009. The specific targeting of guidance receptors within neurons: who directs the directors? Dev Biol, 327:4–11.

10.

Schaefer

, Kabir

, Forscher

. 2002. Filopodia and actin arcs guide the assembly and transport of two populations of microtubules with unique dynamic parameters in neuronal growth cones. J Cell Biol, 158:139–152.

11.

Fraichard

, Chassande

, Bilbaut

, Dehay

, Savatier

, Samarut

. 1995. In vitro differentiation of embryonic stem cells into glial cells and functional neurons. J Cell Sci, 108:3181–3188.

12.

Keller

. 2005. Embryonic stem cell differentiation: emergence of a new era in biology and medicine. Genes Dev, 19:1129–1155.

13.

Ban

, Bonifazi

, Pinato

, Broccard

, Studer

, Torre

, Ruaro

. 2007. Embryonic stem cell-derived neurons form functional networks in vitro. Stem Cells, 25:738–749.

14.

Kondra

, Laishram

, Ban

, Migliorini

, Di Foggia

, Lazzarino

, Torre

, Ruaro

. 2009. Integration of confocal and atomic force microscopy images. J Neurosci Methods, 177:94–107.

15.

Alessandrini

, Facci

. 2005. AFM: a versatile tool in biophysics. Meas Sci Technol, 16:R65–R92.

16.

Parpura

, Haydon

, Henderson

. 1993. Three-dimensional imaging of living neurons and glia with the atomic force microscope. J Cell Sci, 104:427–432.

17.

McNally

. 2004. Imaging and manipulating living neurons with atomic force microscopy. Conf Proc IEEE Eng Med Biol Soc, 6:4473–4476.

18.

Godement

, Wang

, Mason

. 1994. Retinal axon divergence in the optic chiasm: dynamics of growth cone behaviour at the mideline. J Neurosci, 14:7024–7039.

19.

Mason

, Erskine

. 2000. Growth cone form, behavior, and interactions in vivo: retinal axon pathfinding as a model. J Neurobiol, 44:260–270Erratum in: J Neurobiol (2000) 45:134.

20.

Al-Jahdari

, Suzuki

, Yoshida

, Noda

, Shirai

, Saito

, Goto

, Nakano

. 2008. Growth cone collapse and neurite retractions: an approach to examine X-irradiation affects on neuron cells. J Radiat Res (Tokyo), 49:481–489.

21.

Tamagnone

, Comoglio

. 2004. To move or not to move? Semaphorin signalling in cell migration. EMBO Rep, 5:356–361.

22.

Saito

. 1997. Effects of lysophosphatidic acid on primary cultured chick neurons. Neurosci Lett, 229:73–76.

23.

Fuhr

, Richter

, Zimmermann

, Hitzler

, Niehus

, Hagedorn

. 1998. Cell traces—footprints of individual cells during locomotion and adhesion. Biol Chem, 379:1161–1173.

24.

Richter

, Hitzler

, Zimmermann

, Hagedorn

, Fuhr

. 2000. Trace formation during locomotion of L929 mouse fibroblasts continuously recorded by interference reflection microscopy (IRM) Cell Motil Cytoskeleton, 47:38–47.

25.

Zimmermann

, Richter

, Reichle

, Westphal

, Geggier

, Rehn

, Rogaschewski

, Bleiss

, Fuhr

. 2001. Mammalian cell traces: morphology, molecular composition, artificial guidance and biotechnological relevance as a new type of “bionanotube.” Appl Phys, 73:11–26.

26.

Kirfel

, Rigort

, Borm

, Herzog

. 2004. Cell migration: mechanisms of rear detachment and the formation of migration tracks. Eur J Cell Biol, 83:717–724.

27.

Brown

, Bridgman

. 2009. Disruption of the cytoskeleton during Semaphorin 3A induced growth cone collapse correlates with differences in actin organization and associated binding proteins. Dev Neurobiol, 69:633–646.

28.

Kolpak

, Jiang

, Guo

, Standley

, Bellve

, Fogarty

, Bao

. 2009. Negative guidance factor-induced macropinocytosis in the growth cone plays a critical role in repulsive axon turning. J Neurosci, 29:10488–10498.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

1.65 MB

0.00 MB