Increase in Growth Cone Size Correlates with Decrease in Neurite Growth Rate

Several important discoveries in growth cone cell biology were made possible by the use of growth cones derived from cultured Aplysia bag cell neurons, including the characterization of the organization and dynamics of the cytoskeleton. The majority of these Aplysia studies focused on large growth cones induced by poly-L-lysine substrates at early stages in cell culture. Under these conditions, the growth cones are in a steady state with very little net advancement. Here, we offer a comprehensive cellular analysis of the motile behavior of Aplysia growth cones in culture beyond this pausing state. We found that average growth cone size decreased with cell culture time whereas average growth rate increased. This inverse correlation of growth rate and growth cone size was due to the occurrence of large growth cones with a peripheral domain larger than 100 μm2. The large pausing growth cones had central domains that were less consistently aligned with the direction of growth and could be converted into smaller, faster-growing growth cones by addition of a three-dimensional collagen gel. We conclude that the significant lateral expansion of lamellipodia and filopodia as observed during these culture conditions has a negative effect on neurite growth.


Introduction
The neuronal growth cone is the motile tip of the neurite and an important mediator of neural plasticity [1][2][3]. It directs axons and dendrites to their respective targets during neural development and regeneration, by sensing and responding to a multitude of both chemical and physical cues [4,5]. As a result of the extreme sensitivity of growth cones to their environment, dramatic variation in morphology and motility of the same growth cone has often been observed in vivo, reflecting the complex nature of the microenvironment [6][7][8]. Cell culture provides an alternative to study the behavior of growth cones in a defined experimental setting [9][10][11]. Cultured neurons retain the ability to form neurites as well as axons and dendrites [9,10,12,13]. The growth cones formed in vitro are characterized by three domains: (1) the peripheral (P) domain, which includes both filopodial and lamellipodial extensions at the distal edge; (2) the central (C) domain, which is rich in organelles and vesicles and bridges the P domain of growth cone and the newly formed neurite shaft; and (3) the transition (T) zone between P and C domain, often bearing membrane ruffles [1,2,5]. This domain assignment is a general feature of growth cones, whereas the relative size and shape of individual domains for a given growth cone vary greatly depending on species, cell type, and culture condition. In a complete cycle of neurite outgrowth, the protrusion of filopodia and lamellipodia in the P domain of growth cone is followed by the invasion of microtubule bundles, organelles, and vesicles into the P domain (engorgement) and the conversion of the growth cone neck into a new segment of the neurite (consolidation) [5,[14][15][16].
The bag cell neuron from the sea slug Aplysia californica is one of the most extensively used model systems for investigations involving electrophysiology [17][18][19], neuropeptide synthesis and secretion [17,[20][21][22], neuronal motility and guidance [23,24], cytoskeletal dynamics [25][26][27][28][29], and cellular biophysics [30][31][32]. The preference for Aplysia bag cell neurons results in part from the large cell body (∼50 m diameter) and the large growth cone in cell culture, especially when plated on poly-L-lysine-(PLL-) coated coverslips (∼1000 m 2 total growth cone surface area). Comparatively, vertebrate neuronal growth cones are typically smaller than 100 m 2 . This 10x increase in size is advantageous not only for quantitative high-resolution analysis of numerous growth cone activities [27,28,33,34], but also for various biophysical manipulations [29,30]. Not surprisingly, several key findings regarding growth cone morphology and dynamics were made in studies using Aplysia bag cell neuronal growth cones, thanks to the advantages provided by both the large size and the stereotypic domain organization [23-25, 30, 35]. In addition, the organized geometry of the Aplysia growth cone is an attractive target for modeling of growth cone dynamics and motility [36]. Despite the pivotal role that Aplysia bag cell neuronal growth cone played in cellular neurobiology, a basic description of growth cone behavior in vivo has been missing due to challenges in breeding and imaging developing Aplysia [37,38].
In previous in vitro studies involving Aplysia bag cell neurons, no staging of growth cone development has been performed as it has been done for hippocampal neurons [10] or Aplysia buccal ganglion cell neurons developing normally [14] or after axotomy [39]. Instead, almost all previous studies involving Aplysia bag cell growth cones have focused on large growth cones on PLL substrates and short-term motility at steady state, during which neither significant growth nor retraction occurred. Accounts of bag cell growth cones showing significant advancement are rare, and the growth rate was usually lower than that of other types of growth cones, unless neurons were stimulated by apCAM substrates [24,30,33,34] or plated on hemolymph with or without laminin [31,40]. It is unclear whether the pausing state of large growth cones is due to intrinsic properties or due to limiting factors presented by the culture environment or both.
In this study, we provide a detailed analysis of the behavior of Aplysia bag cell growth cones beyond the large, pausing state typically observed in culture. We show that average size of growth cone decreased with increasing cell culture time, whereas average growth rate increased. We found that this inverse correlation was due to the occurrence of large growth cones (with P domains larger than 100 m 2 ) at early stages of cell culture and that large growth cones usually had minimal growth rate (less than 2 m/h). C domains of large growth cones were less aligned with the direction of growth when compared to C domains of smaller, fast-growing cones. Large growth cones were formed on two-dimensional (2D) PLL substrates by lamellipodial veils connecting numerous filopodia. Furthermore, we found that large growth cones could exit the pausing state either by branching or by application of a type I collagen gel on top of the growth cone. These findings suggest that widely spread growth cones as found on 2D PLL substrates are not conducive to fast neurite growth.

2.2.
Imaging. Live cell imaging was performed at room temperature in L15-ASW on a Nikon TE2000 E2 inverted microscope (Nikon, Melville, NY) with 60x oil immersion objective lens, with additional 1.5x magnification. Focus during long time-lapse sequences was maintained with the Nikon Perfect Focus system. Images were collected with a Cascade II charge-coupled device camera (Photometrics, Tucson, AZ) controlled by MetaMorph 7.8 software (Molecular Devices, Sunnyvale, CA). Imaging chamber was assembled as previously reported for cells cultured on coverslip [41], and fresh L15-ASW was supplemented for imaging longer than 1 h. For cells cultured on glass-bottom dish, fresh medium was exchanged every hour. Time-lapse series were acquired at 30 s intervals for 3.5 h (Figure 4

Neurite Growth Rate and Growth Cone Size Analysis.
Neurites were identified and imaged at different time points after cell plating: between 24 and 27 h; between 48 and 51 h; and between 72 and 75 h after cell plating. MetaMorph 7.8 software was used for image analysis. To determine individual neurite growth rates, the displacement of a neurite tip over a three-hour interval (3.5 h for Figure 2) was measured and divided by the time. P domain size was measured and used to correlate with the growth rate of an individual growth cone. To account for different shape and domain organization of big and small growth cones, only the P domain directly in front of C domain was included. Specifically, the center of the C domain was determined, which served as the origin of a Cartesian coordinate system superimposed on the growth cone with the -axis in the direction of growth cone advance. Only the area of the P domain that was in front of the -axis was included in the size measurements.

2D-3D Transition
Culture. Cells were cultured on PLLcoated coverslips for 24 h to allow for the formation of growth cones, and then an imaging chamber was assembled. Liquid 2 mg/mL type I rat tail collagen (Corning, Corning, NY) based gel mixture (containing 25% Aplysia hemolymph [40] and 20 g/mL PLL) was introduced into the chamber on top of the cells and allowed to solidify at 37 ∘ C for 1 h. The imaging chamber was placed back into dishes with L15-ASW for continued cell culture at RT, until chambers were mounted for imaging. Cells cultured in this setup maintain a healthy morphology for at least 10 d.

Statistics.
Graphs were made in GraphPad Prism 6 (GraphPad Software, Inc., La Jolla, CA). Mann-Whitney test ( Figures 1 and 6) or Dunn's test (Figure 3) was used following normality testing to compare the P domain sizes and growth rates of different groups. Since P domain and growth cone size data was not normally distributed and varied over a wide range, we used a log scale and whisker plots with the box representing the 25th and 75th percentile, middle line indicating the median, and whiskers representing the min and max values. Significance was established at < 0.05.

Aplysia Growth Cones Become Smaller with Time in Cul-
ture. Previous studies on Aplysia bag cell neuronal growth cones have mainly focused on large fan-shaped specimens cultured for one or two days on PLL substrates [24,27,28,33,34]. Such growth cones exhibit not only distinct organization of cytoplasmic domains but also a highly organized actin and microtubule cytoskeleton with well-characterized dynamic properties. However, the vast majority of these growth cones do not significantly translocate during a typical observation time of 30 to 60 min. Here, we compared the sizes of growth cones from cultured bag cell neuron on PLL-coated coverslips at different times in culture. At 18 h after plating, multiple large growth cones with a diameter between 50 and 100 m appeared around the cell body. These growth cones share stereotypic organization, with a flat P domain and organellerich C domain separated by the T zone that features extensive membrane ruffles (Figure 1(a)). By 78 h, growth cones moved further away from the cell body and were considerably smaller in size. The smaller growth cones had fewer filopodia and a less obvious T zone (Figure 1(a)). The portion of the filopodia extending beyond the leading edge became proportionally longer compared to the portion of the filopodia embedded in the lamellipodia (arrow in the lower right image). Statistical analysis revealed a wide distribution of P domain size at both time points (note the log scale), as well as a significant drop of the average P domain size at 78 h when compared to the 18 h time point (Figure 1(b)).

Large Growth Cones Tend to Have Lower Growth Rates.
In order to test if growth cone size has an impact on growth rate, we took images of individual growth cones on the second day of plating, at 3.5 h interval (at 53 h and 56.5 h after plating). As can be seen from Figure 2(a), the large growth cone exhibited minimal advancement within 3.5 h, while the two smaller growth cones translocated 45 and 50 m, respectively, over the same period of time. A plot of growth cone advance rate versus P domain size indicated a nonlinear correlation, with the highest growth rate at around 50 m 2 and a decline of growth rate for growth cones with P domain larger than 100 m 2 (Figure 2(b)). Growth cones with P domains smaller than 100 m 2 did not show a clear correlation between size and advance rate (inset in Figure 2(b)). Growth cones with P domains larger than 200 m 2 were less abundant and usually had growth rates of less than 2 m/h. In conclusion, our results show that large growth cones tend to grow slowly, whereas small growth cones do not have a clear correlation between size and advance rate.

Neurite Growth Rate Increases with Time in Culture as
Growth Cone Size Decreases. Prompted by the findings that growth cone size decreases with increased cell culture time and that larger growth cones exhibit lower growth rates, we set out to test whether an increase in growth rate can be observed at later time points of cell culture. To this end, images of individual growth cones were taken at 3 h intervals at 24, 48, and 72 h after cell plating (Figure 3(a)). As predicted, we observed a significant decrease in average P domain size, concomitant with a significant increase in average growth rate with time in culture (Figures 3(b) and 3(c)). Plots of growth rate versus P domain size for individual growth cones, however, showed a nonlinear correlation (Figure 3(d)) in agreement with data shown in Figure 2(b). These plots revealed that the P domain size is a poor predictor of growth rate for individual growth cones with P domain size smaller than 50 m 2 ; however, it becomes a better predictor for larger growth cones, since 100% of growth cones with P domain size larger than 100 m 2 have small growth rate (less than 2 m/h). In summary, the decreased percentage of large growth cones at later time points of cell culture accounts for the increase in average neurite growth rate with time in culture. At later times in culture, the majority of neurite outgrowth was carried out by smaller growth cones with P domain size of less than 50 m 2 .

Large and Small Growth Cones Show Different C Domain and T Zone Dynamics.
In order to identify other differences between large and small growth cones besides size that might explain their different growth rates, we used highresolution live cell imaging over the time course of 3 h to study their motile behavior. In agreement with previous findings (Figure 2), growth cones with P domains larger than 100 m 2 rarely showed net advancement (Figure 4(a); Video 1 in Supplementary Material available online at http://dx.doi .org/10.1155/2016/3497901). These large, fan-shaped growth cones exhibited swaying of the C domain that terminates at membrane ruffles-also referred to as intrapodia-in the T zone [42] and a gradual decrease in the extent of membrane ruffles in the T zone over time. In contrast to the large growth cones, small, motile growth cones advanced by protruding lamellipodia between newly formed filopodia, while avoiding the formation of a large fan-shaped P domain (Figure 4(a), small growth cone on the right; Figure 4(b); Video 2). Furthermore, in smaller growth cones, the C domain followed more consistently the base of filopodia and maintained its directionality better than in large growth cones (Figure 4(b); Video 2). We measured the angle between the mean orientation of the C domain boundary and the direction of growth cone advancement as an indicator for the persistency of the C domain advancement and engorgement (Figure 4(c)). We found that large, pausing growth cones frequently showed swaying of the mean C domain orientation along with splitting of the C domain, suggesting that persistent C domain advance could be more challenging for large growth cones (Figure 4(c)). On the other hand, small advancing growth cones exhibited a more persistent orientation of the C domain. After spending a significant amount of time (typically more than 3 h) in the pausing state, a large growth cone either completely retracted or retracted some of its P domain to form smaller growth cones. In this way, large growth cones gave rise to new neurites through branching from their original P domains (Figure 4(d)). In summary, the persistent advance of the C domain is more common in small growth cones and may contribute to the faster growth of small growth cones when compared to large fan-shaped growth cones. characterize the cellular events that lead to the formation of large growth cones, which is shown in Figure 5  state and became completely retracted within the next 10 h (data not shown), smaller GC2 started to protrude again after formation of membrane ruffles both along the leading edge and in the T zone around 9 h and by elongation of filopodia from the P domain edge between 9 and 10 h ( Figure 5(c); Video 3). Thus, the formation of flat P domain and T zone seems to be a hallmark of large, pausing growth cones. Furthermore, these observations suggest that local changes in actin and membrane dynamics need to occur in order to exit this pausing state.

Collagen I-Based Gel Promotes Faster Outgrowth of Bag Cell Neuronal Growth Cones.
Our results so far suggested that the formation of large, fan-shaped growth cones after 1 day in culture was caused by a significant expansion of filopodia and lamellipodia, which are promoted by  the two-dimensional (2D) PLL substrate. In an attempt to rescue large growth cones from this pausing state, we provided them with an additional substrate support from the dorsal side. In doing so, we designed an anisotropic threedimensional (3D) collagen culture system. After growth cones had formed on PLL-coated coverslips, a layer of a collagen I-based gel containing 25% Aplysia hemolymph [40] and 20 g/mL PLL was introduced on top of the neurons, and the cells were cultured for an additional 24 h (Figure 6(a)). This 2D-3D transition in culture environment had a clear impact on the morphology of growth cones and promoted neurite outgrowth when compared to growth on 2D PLL substrates ( Figure 6(b)). Figure 6(c) shows a large growth cone on PLL before and after introduction of the collagen gel. Compared with the growth cone on plain PLL substrates, growth cones in 3D culture were consistently smaller and had less distinct domain separation (Figures 6(c) and 6(d)). Average total growth cone size before gel application was 1238 ± 178 m 2 ( = 37) and after gel application was 70 ± 24 m 2 ( = 25; Figure 6(e)). Time-lapse imaging of growth cones dynamics advancing at the 2D-3D interface revealed mean growth rates of 13.4 m/h (Figures 6(f) and 6(g); Video 4). Addition of hemolymph to the collagen gel was critical to convert the large pausing growth cones to smaller advancing growth cones. However, we do not believe that the changes observed were solely due to hemolymph, since growth cones cultured on planar surfaces coated with PLL and hemolymph result in a morphological phenotype that is midway between growth cones grown on PLL and at the 2D-3D interface [40]. The neurites formed after gel addition were more rounded and smooth, a sign of neurite consolidation (Figures 6(d) and 6(g); Video 4). These rounded neurites induced by the addition of 3D collagen gel were more similar to the preexisting neurite following dissection of the abdominal ganglion ( Figure 5(a)) and different from the more spread neurites generated on the plain PLL substrates (Figure 4). Therefore, application of the collagen gel promoted neurite outgrowth of bag cell neurons, possibly by providing substrate support on the dorsal side of the growth cone and triggering smaller growth cones.

Stages of Neurite Outgrowth of Cultured Aplysia Bag
Cell Neurons. Growth cones formed by Aplysia bag cell neuron cultured on PLL-coated solid surfaces have proven to be an excellent model system for the past 30 years, as it enabled seminal discoveries regarding the nature of F-actin retrograde flow [23,35], substrate-cytoskeleton coupling [24,32,43], microtubule and actin dynamics [27,28,33,34], and cellular mechanotransduction [30][31][32]. Common features of the growth cones analyzed in the aforementioned studies are large size (typically with P domains larger than 100 m 2 ), fan-shape, and clear domain distinction, the combination of which provides unique advantages especially for highresolution measurements of cytoskeletal dynamics and cellular mechanics. While being a well-defined system, these large growth cones represent only one type among the diverse growth cone morphologies observed in vitro and in vivo. It is therefore necessary to investigate the motile behavior of bag cell growth cones beyond the typically reported shape to better compare findings made with different growth cone model systems. We believe that the basic mechanisms of how cytoskeletal dynamics is harnessed for growth cone motility and advance are highly conserved across different species and growth conditions. What varies among different growth cone systems, growth states, and conditions is the relative contribution of individual processes to net outgrowth, such as (1) actin assembly and recycling, (2) myosin activity, (3) adhesion and clutching, and (4) membrane delivery and recycling. Thus, the large, fan-shape Aplysia growth cone can be considered a transient "pausing growth cone," in which actin assembly and retrograde flow rates are well balanced [27]. It can exit this pausing stage through a new stimulus such as an adhesion protein-coated bead or needle [24,30], by branching or introduction of a collagen gel (this study) as further discussed below.
To achieve a more comprehensive assessment of the motile behavior of cultured bag cell growth cones, we analyzed the morphology and motility of bag cell growth cones on PLL-coated coverslip at different time points in culture in an unbiased manner with respect to growth cone morphology and size and describe how the large growth cones develop over time. During the first 24 h after plating, multiple large, fan-shaped growth cones formed with P domains larger than 100 m 2 . These growth cones formed from preexisting neurites or de novo from the cell body. The large P domain was induced by multiple lamellipodial veils protruding and connecting an increasing number of filopodia ( Figure 5(a); Video 3). The formation of these large growth cones took about 3-5 h, and they typically spent a significant amount of time (more than 3 h) in the pausing state (Figure 4(a); Video 1). Large growth cones then typically branched into neurites with growth cones of smaller size between 24 and 72 h after plating (Figures 1, 3, and 4(d)). None of the neurites was morphologically distinct from others indicating that an axon cannot be distinguished from dendrites in bag cell neurites, which is not surprising considering that these cells primarily have neurosecretory function. Although a simple linear correlation between P domain size and growth rate for individual growth cones could not be established for all observed growth cones, a lower growth rate (smaller than 2 m/h) was usually found for large growth cones (P domain size more than 100 m 2 ), which is in agreement with previous study where instantaneous growth rates of large growth cones were measured (Figure 2(b)) [40]. Small growth cones can advance as fast as 10 m/h on PLL substrates in the absence of any additional growth promoting factors (Figure 2(a); Video 2), suggesting the lower growth rate seen for large growth cones is likely due to the significant growth cone enlargement. When neurite outgrowth was promoted by either adding of hemolymph or plating on coverslips coated with hemolymph/laminin, a reduction in growth cone size could be readily observed [40,44]. Bag cell neurons plated on laminin/ hemolymph-coated silicone gel also displayed smaller size and faster growth rate [31]. Thus, a smaller size seems to be a prerequisite for fast growth in the case of Aplysia bag cell neuronal growth cones, whereas the large growth cones conventionally used for imaging cytoskeletal dynamics are representative of a pausing state. Because of this, one should therefore be careful when directly comparing absolute numbers such as retrograde flow rates gained from pausing Aplysia bag cell neuronal growth cones with faster-growing growth cones derived from other species or when extrapolating to growth cones in vivo. It will be interesting to investigate whether actin assembly and retrograde flow rates are different between large, pausing and small, advancing growth cones, in order to better understand the causal relationship between growth cone size and advance rate. A definite causal relationship between size and growth rate, however, could only be established if we would find a substrate-independent way to directly enlarge the size of fast moving Aplysia bag cell neuron growth cones.

Large and Small Growth Cone Have Different C Domain
Dynamics and Ruffling Activity. Whereas the dynamics of large Aplysia bag cell growth cone has been extensively studied on shorter time scales, we have complemented this knowledge here with live cell imaging on longer time scales. Two cellular features were discovered when large growth cones developed from small growth cones: (1) the C domain frequently changed its direction when large growth cones were in a pausing state and sometimes split into two domains; (2) T zone ruffles appeared soon after the formation of a large growth cone and then gradually disappeared (Figure 4(a); Videos 1 and 3). Efficient engorgement of the C domain into the area where the P domain was previously located is an important step during the formation of axons [14]. Swaying of the C domain as observed in large growth cones (Figure 4; Video 1) may indicate a decision-making phase before final engorgement occurs in a particular direction. The varying angle between C domain orientation and the direction of growth in large growth cones (Figure 4(c)) could be caused by the symmetrically expanding P domain and the fact that there are not enough microtubules and vesicles to fill the expanding C domain evenly. When the symmetry in the P domain was broken by disruption of actin bundles, large growth cones could be reoriented [45]. Splitting of the C domain was usually followed by branching of growth cones. It appears that large growth cones can only exit the pausing state by branching into smaller ones, or they completely retract. Branching happened either spontaneously (Figure 4) or by applying a collagen gel (Figure 6(d)). A similar branching pattern was observed after treatment of Helisoma growth cones with conditioning factors [46]. In contrast, in fast-growing small growth cones, the P domain was biased towards the direction of a small number of filopodia, and C domain maintained better alignment with the direction of growth (Figures 4(b) and 4(c); Video 2). Interestingly, in small growth cones, ruffles in the P domain were often generated towards the direction of advancement along the leading edge, preceding the formation of filopodia from P domain (Figures 4(b) and 5(c); Video 3), whereas, in large growth cones, ruffles were usually confined to the T zone and eventually disappeared (Figure 4(a)). Whether the ruffles along the leading edge of small growth cones and in the T zone of large growth cones have the same function with respect to motility is unclear at this point. We speculate that the initial ruffling activity in the T zone of large growth cones may indicate an attempt of the growth cone to grow vertically as opposed to expanding horizontally.

Rescue of Growth Cone Translocation through 3D Collagen
Gel. In vivo data suggests a correlation between growth cone morphology and motility: rapidly growing retinal ganglion cell growth cones in the optic nerve are smaller and have simple shape, whereas, at decision-making positions such as the optic chiasm, growth cones adopt a more elaborate fan-shape with filopodial and lamellipodial extensions and grow at a significantly reduced rate [7,8]. The large growth cones from Aplysia bag cell neurons cultured on PLL may behave similarly to the second group, searching for chemical and physical cues that are scarce under the current culture conditions. For Aplysia bag cell neuron, this pausing state usually takes a long time (typically longer than 3 h) and redirecting its growth has been difficult. Physical interaction with either an Aplysia cell adhesion molecule-(apCAM-) coated bead or microneedle could lead to reorientation of C domain and increased advance rate [24,30]. However, since the microneedle was kept in the same position in these experiments, increased neurite growth rate did not persist once the C domain reached the microneedle. Through application of collagen I-based gel containing hemolymph on top of the growth cones, we were able to create and maintain a topographical cue that could reliably rescue the forward translocation of large growth cones ( Figure 6; Video 4). Growth cones found in this culture environment had smaller size, faster growth rate, and a dramatically different morphology that was reminiscent of growth cone on patterned substrate [47]. The neurons in this 2D-3D transition setup are not completely independent from the planar PLL-surface as indicated by the fact that neurites tend to remain close to the cover glass surface. To our knowledge, this 3D collagen gel is the only culture method that can induce well-consolidated neurites from Aplysia bag cell neurons. We believe the growth cones formed in this 2D-3D transition setup may be more similar to the growth cones in living animals with respect to morphology and motility. The relative contributions of individual processes of cytoskeletal and membrane dynamics may differ between the 2D and 3D environments resulting in the observed differences in morphology and motility. The effect of this collagen gel-based approach may be attributable to the presence of micron scale topographical cue that can be sensed and responded to by the growth cone. By creating and maintaining this local difference, individual growth cones maintain a rapidly advancing state similar to retinal ganglion cell growth cones in the optical tract. Although technically challenging, it will be interesting to quantitatively analyze motility, cytoskeletal dynamics, and force production of bag cell growth cones in 3D and compare with existing 2D data to better understand the machinery that powers growth cone advancement in vivo.

Conclusion
Here, we have shown that the typically observed large fanshaped Aplysia bag cell growth cones formed on PLL-coated coverslip represent a specific pausing state early during cell culture, which is caused by a significant expansion of the lamellipodia and filopodia in the P domain. Small growth cones on PLL often grow faster than large growth cones, although a simple correlation between size and advance rate could not be established for small growth cones on PLL. An inverse correlation between growth cone size and growth rate, however, was found for large growth cones. The large pausing growth cones exhibited poor alignment between C domain and the direction of advancement, which could be indicative of the growth cone sampling the environment. In contrast, the small rapidly growing growth cones generated ruffles at the distal edge of P domain and grew as fast as 10 m/h without additional growth promoting factors. Application of collagen I-based gel promoted faster growth of bag cell growth cones and converted them into a more in vivo-like morphology. We conclude that significant lateral expansion of lamellipodia and filopodia as observed on PLL substrates during the initial culture period negatively affects neurite growth.