Neurons were cultured on silicone elastic membranes (Specialty Manufacturing Inc., Saginaw, MI) and deformed with the application of a pressure-pulse of air (Magou et al., 2011; Figure 1). This rapid deformation causes the stretching of neurons that are adherent to the substrate. In this study, we applied a uniaxial stretch to isolated axon tracts spanning a 2 mm cell free gap. In order to injure axons crossing the cell free gap, a rigid mask printed from acrylonitrile butadiene styrene (ABS) plastic with a 2-mm gap in the center was placed underneath the well prior to injury. Masks were printed using a three-dimensional rapid prototyping system (SST 1200es, Dimension, Inc., Eden Prairie, MN) designed using Pro-Engineer software (PTC, Needham, MA).

Uniaxial stretch injury to axons was measured in terms of strain, defined by the amount the membrane is stretched divided by its original length. The rate of injury was measured in terms of the strain rate, defined by the applied strain over the rise time to maximum strain (i.e., a 60% strain applied over a period of 20 ms has a strain rate of 30 s⁻¹). Axons were injured at uniaxial strains of 40 or 60% at a common strain rate of 30 s⁻¹. These injury parameters have been experimentally determined to induce the formation of undulations immediately post injury and axonal swellings within 2 h after injury in cortical axons. Both cortical and DRGN axons were injured between 10 and 12 days in vitro (DIV) at both strains.

Custom polyetheretherketone (PEEK) wells assembled with elastic silicone membranes were submerged in dH₂O and autoclaved for 1 h. For cortical cultures, wells were coated with poly-L-lysine (0.05 mg/ml; Peptides International, Louisville, KY) and incubated at 37°C overnight for optimal adsorption into the surface of the silicone. On the day of isolation, wells were rinsed three times with sterile dH₂O and left to dry prior to plating. For DRGN cultures, wells were coated with 10 μg/ml high molecular weight poly-D-lysine (BD Biosciences, Bedford, MA) in phosphate buffered saline (PBS) for 1 h. Wells were then rinsed three times and coated with Matrigel (1:35 dilution, BD Biosciences, Bedford, MA) extracellular matrix overnight, then aspirated and allowed to dry the following day prior to plating.

In order to create a cell free zone for uniaxial stretching of isolated axons, silicone sheeting was cut into strips of 2 mm in width to fit the size of the gap in injury masks. After drying, 2 mm wide silicone strips were placed at the center of each well using a sterilized deformation mask for placement guidance. On the day after plating cells (DIV 1), strips were removed to create a cell free zone. Developing axons traverse the gap creating isolated axons for stretch injury. Cortical axons completely integrate across the 2 mm gap within 10 DIV (Iwata et al., 2004; Monnerie et al., 2010). Here it was found that DRGN axons also completely integrate across the gap by 10 DIV, the time at which both cultures were injured.

For cortical cultures, cortices were isolated from E17 Sprague-Dawley rat pups (Kingston, NY), the meninges were removed and cortices were stored on ice in Hanks Balanced Buffer Solution (HBSS) containing Ca²⁺ and Mg²⁺ followed by a brief calcium switch in Ca²⁺ and Mg²⁺ free HBSS. Tissue was then digested in 0.025% Trypsin containing DNase I (1.0 mg/ml) and EDTA (0.2 g/L) for 30 min in the incubator at 37°C. Trypsin solution was then removed and neutralized using 10% fetal bovine serum (FBS) in Ca²⁺ and Mg²⁺ free HBSS. After removal of FBS, cells were suspended in 500 μl of cortical growth medium [Neurobasal media, 2% B-27, 1% penicillin-streptomycin, and 0.4 mM L-glutamine (Thermo Fisher Scientific, Waltham, MA)]. Cells in the growth medium were then centrifuged at 2,000 RPM for 5 min. The supernatant was removed, and cells were resuspended in 500 μl of media. After the addition of 5 ml growth media, cells were filtered through a sterilized nylon mesh of 100 μm pore size, followed by 50 μm. Cells were then counted and plated at a density of 375,000 cells/cm².

For DRGN cultures, spinal cords were removed from E17 Sprague-Dawley rat pups and stored in ice cold Leibovitz’s L-15 medium. DRGNs were removed from the spinal cord, collected, and incubated in trypsin for 1 h at 37°C. Following tissue digestion, trypsin was removed and replaced with 100% FBS and centrifuged at 6,000 RPM for 10 min. The supernatant was removed, and cells were resuspended in a growth medium (Neurobasal medium, 2% B-27, 1% penicillin-streptomycin, 0.4 mM L-glutamine, 500 μl 20% D-glucose, 20 ng/ml nerve growth factor, 20 μM 5-fluoro-2’-deoxyuridine + 20 μM Uridine). Cells were plated within 100 μl droplets on both sides of the silicone strip and allowed to attach for 1 h. After 1 h, an additional 500 μl of growth media was added to each well. Media was changed after 24 h and every three days following. After 3 days in culture, cells were maintained in growth media without mitotic inhibitors. Media was changed 24 h after isolation and again every three days following. All animal work conforms to the National Research Council’s Guide for the Care and Use of Laboratory Animals and protocols were approved by the Rutgers-University Newark Animal Institutional Animal Care and Use Committee.

Axonal morphological alterations in the region of injury were analyzed in two ways: immediate formation of undulations and the time frame for recovery to the original axon length. Phase contrast time-lapse imaging was used to record changes in axonal length over 20 min after injury with one picture taken every min of the recording. Undulatory deformations were characterized immediately after stretch and quantified by percent strain:

where L₀ is the original length of the segment, and L(t) is the length of the segment at time ‘t’. Percent recovery was determined similarly:

where 100% would represent the complete recovery of strain to the original axon length L₀.

To assess degeneration, a region of interest was selected within the unaxially stretched area of isolated axons prior to the start of the experiments. Images were captured prior to injury (0 h) and at 1, 2, 4, and 24 h post injury. At each time point, images within the region of interest were visually compared to the corresponding photo taken before injury. Images were assessed based on the presence or absence of the following characteristics: the development of axonal swellings along the length of injured axons; and degenerating axons, characterized by visible disconnection of a portion of the axon or complete loss of the axon within the field of view (Smith and Meaney, 2000). Six DRGN and eight cortical cultures from three different tissue preparations injured at 60% strain as well as eleven DRGN and six cortical cultures from three additional preparations injured at 40% strain were included in morphological analyses.

Changes in intracellular calcium were measured from injured cortical and DRGN axons. Cells were gently rinsed three times with HBSS containing Ca²⁺ and Mg²⁺ and loaded with 4 μM Fluo-4 AM (F-14201, Invitrogen) solubilized in 100% anhydrous dimethylsulfoxide (DMSO) with Pluronic F-127 cellular detergent (0.01% w/v). Stock dye was diluted in a controlled saline solution (CSS; 120 mM NaCl, 5.4 mM KCl, 0.8 mM MgCl₂, 1.8 mM CaCl₂, 15 mM glucose, 25 mM HEPES at pH 7.4; Smith et al., 1999). Cells were incubated in loading solution for 30 min. The loading solution was then removed and replaced with fresh CSS to promote de-esterification of the dye. Prior to injury, cultures were rinsed one time with CSS for 3 min.

Fluorescent intensity was captured using a Nikon Eclipse TE2000 inverted microscope with a 20× objective in conjunction with a 75 watt xenon arc lamp and Photometrics CoolSNAP EZ CCD Camera. All camera settings were adjusted at the beginning of the experiment during the baseline measurement and not adjusted further. To minimize photobleaching of cells and increase light sensitivity, 2 × 2 camera binning was used. Automatic image acquisition was programmed using NIS Elements software (Nikon Instruments, Inc, Melville, NY) and controlled the shutter (Lambda SC Smart Shutter, Sutter Instruments, Novato, CA). Images were collected at 2 s intervals for thirty seconds of baseline measurements prior to injury followed by 2 min of recording post-injury and then at 1-min intervals for the following 5 min.

Fluorescent intensity was measured in individual axons using ImageJ software. The “StackReg” plugin was used to correct displacement between frames during and after injury (Thévenaz et al., 1998). Individual axons were outlined using freehand selection in ImageJ and mean gray value measurements were taken from each selected region (each axon) in each frame of the recording. Three measurements of background intensity were taken in each frame of the recordings, averaged, and subtracted from the intensity measurements (Schneider et al., 2012). Changes in fluorescent intensity were reported in terms of the ratio of the measured fluorescence at each time point (F) divided by the mean baseline fluorescence (F₀), where baseline intensity by definition has a value of F/F₀ = 1.

The dose for lidocaine blockage of sodium channels in DRGNs was determined experimentally based on previous findings (Roy and Narahashi, 1992). Cultures were pre-treated with lidocaine after loading and rinsing of Fluo-4 calcium indicator dye and before stretch injury. We found that a 200 μM concentration of lidocaine was sufficient to block both TTX sensitive and resistant sodium channels under injury of 60% strain at a rate of 30 s⁻¹. To verify that extracellular calcium enters the axon, cultures were injured using HBSS free of Ca²⁺ and Mg²⁺ to rinse calcium and magnesium from the extracellular environment.

At 2 h post-injury, DRGN cultures were rinsed with phosphate-buffered solution (PBS) and fixed for 20 min in 4% paraformaldehyde. Axons were then permeabilized with 0.3% Triton-X, blocked with 3% bovine serum albumin, and immunostained with a rabbit anti-neurofilament (NF200; 1:500; Sigma-Aldrich, St. Louis, MO), rabbit polyclonal anti-NaCh brain type II (epitope corresponds to the I-II intracellular loop; 1:50; Sigma-Aldrich, St. Louis, MO), and rabbit polyclonal anti-NaCh pan (epitope corresponds to the III-IV loop; Nav1, 1:100; Sigma-Aldrich, St. Louis, MO). The next day, samples were rinsed three times in 5 min intervals with phosphate buffered saline (PBS) before incubation in Alexa 488 or 594 secondary antibodies for 1 h. Samples were then rinsed three times in 5 min intervals before cutting silicone from the wells and mounting on glass slides.

Data expressed as the mean ± the standard error of the mean. SPSS (IBM, Armonk, NY) statistical software was used for analysis. Differences between the means of cortical and DRGN morphological features were analyzed by the student’s T-test. Calcium measurement data were analyzed with the Kruskal-Wallis nonparametric test due to the data not meeting the assumptions of normality or homogeneity of variances using Levene’s test. Post hoc pairwise comparison was carried out between groups with significance values adjusted by the Bonferroni correction for multiple tests.

To quantify the stretch injury-induced lengthening of axons, DRGN, and cortical axons were uniaxially stretch injured with 60% strain at a rate of 30 s⁻¹ (n = 15 axons). Cultures were labeled with Fluo-4 for fluorescent imaging of axons to enhance contrast. Image J was used to measure axonal lengths prior to injury and immediately after the injury. Cortical axons elongated by 14.14% ± 1.96 over their original lengths. Similarly, DRGNs axons responded to injury with an average elongation of 14.45% ± 1.38 (Figure 2A). The ability of axons to recover to their original length was quantified at 20 min post injury. Cortical axons recovered to 98.35% ± 4.36 of their original length and similarly, DRGNs axons recovered to 97.41 ± 3.23% of their original length (Figures 2B,C). There was no significant difference between the two populations (p > 0.05 for elongation and recovery).

Stretch injury severity scales with strain in cultured cortical neurons (Tang-Schomer et al., 2010). To determine how strain magnitude affects morphological outcome in these two neuronal subtypes, cortical and DRGN axons were stretch injured at strains of 40% and 60% at a common strain rate of 30 s⁻¹. Axonal swellings have been observed along the length of injured axons in previously described in vitro stretch injury models as well as in vivo models of TBI. By 2 h post injury, all DRGN and cortical cultures injured at a strain of 60% showed evidence of axonal swellings (Figure 3). Conversely, less than 20% of the DRGN cultures and 50% of the cortical cultures injured at 40% strain developed axonal swellings. Notably, these mildly injured axons that develop undulations did not always develop axonal swellings, suggesting some axons avoid degeneration.

By 4 h post injury, all DRGN and cortical cultures injured at 60% strain displayed axonal degeneration that progressed in severity through the 24-h period of observation. For injuries at 40% strain, only the DRGN cultures that previously developed swellings had observable axonal disconnection by the 4 h post injury time point. 33% of the cortical cultures injured at a strain of 40% were found to have disconnected axons by 24 h post injury, but none of the cultures in the study did at the 4 h time point. This suggests that there may be a delay in this response in cortical axons injured at a lower strain.

Cortical and DRGN axons were injured to induce an influx of intracellular calcium, a well-established mechanism in cortical neurons that has been reported to scale with the degree of stretch applied (Lusardi et al., 2004). To test whether this occurs in DRGN, both populations were injured at a strain of 60% at a rate of 30 s⁻¹ and a 40% strain at a rate of 30 s⁻¹, and the influx of calcium was measured via the change in Fluo-4 fluorescence signal (Figure 4).

Stretch injured cortical axons responded with an increase in Fluo-4 fluorescence to a mean F/F₀ of 1.68 ± 0.75 (n = 25) at 60% strain (Figure 4A). Injured DRGN axons also elicited an increase in Fluo-4 fluorescence to a mean F/F₀ of 2.20 ± 0.33 (n = 25) at 60% strain. To confirm that the rise in intracellular calcium was due to an influx of extracellular calcium, cultures were injured in a calcium-free buffer. Both cortical and DRGN axons injured without the presence of extracellular calcium did not elicit an intracellular rise in calcium, with mean F/F₀ of 0.99 ± 0.02 (n = 19) and 1.03 ± 0.06 (n = 9) respectively. Post hoc pairwise comparisons indicated that stretch induced calcium influx in DRGN and cortical axons were not significantly different at 60% strain (p > 0.05), but were significantly different from controls excluding extracellular calcium (p < 0.05).

Blocking sodium channels with tetrodotoxin (TTX) in cortical axons prevents calcium entry after injury (Wolf et al., 2001). Unlike cortical neurons, DRGN have TTX resistant sodium channels. To test the mechanism of sodium channel blockage on calcium influx in DRGNs axons, we stretch injured axons at 60% strain, 30 s⁻¹ strain rate in the presence of lidocaine (Roy and Narahashi, 1992). DRGN cultures that received lidocaine treatment prior to injury (n = 18) did not respond with an influx of calcium (F/F₀ = 0.93 ± 0.03; Figure 4A). Lidocaine treatment leads to calcium measurements similar to controls (p = 1 with Bonferroni adjustment) and significantly different from untreated DRGN axons (p < 0.05).

To investigate whether calcium influx into axons scaled with the level of injury, cortical and DRGN axons injured at 60% strain were compared to axons injured at 40% strain. Stretch injured cortical axons responded with an increase in Fluo-4 fluorescence to a mean F/F₀ of 1.68 ± 0.14 (n = 36) at 40% strain, 0.88 ± 0.04 (n = 11) in buffer without calcium (Figure 4B). Injury to DRGN axons also elicited an increase in Fluo-4 fluorescence to a mean F/F₀ of 1.72 ± 0.11 (n = 44) at 40% strain, 0.89 ± 0.07 (n = 15) in buffer without calcium. DRGN and cortical groups injured at 60% and 40% strain were not significantly different (p > 0.05). This data also indicates that calcium influx levels in both DRGN and cortical axons do not scale with the degree of stretch.

The injury induced calcium influx activates proteases that cleave the III-IV intracellular loop of the sodium channel while the I-II loop remains in the membrane of cortical axons (Iwata et al., 2004). Injured DRGN axons were found to respond similarly where the NaCh I-II loop was retained in injured axons and expression of the III-IV is undetectable within the first 20 min after injury (Figure 5). Pre-treatment with either lidocaine or broad-spectrum protease inhibitors prevented the cleavage of the intracellular III-IV loop and retained after injury in DRGN axons (Figure 6).