Na_v1.7-immunofluorescence was observed in many nerve fibers identified by N52/PGP9.5 staining in all regions of both normal and painful dental pulp samples, including fibers within the pulp horn (Fig. 1) and axon bundles located throughout the pulp (Fig. 2). Fibers with Na_v1.7 staining most commonly included those with both N52 and PGP9.5 (Figs. 1A–D and 2A,C). The N52/PGP9.5 staining (and especially the N52 staining) was mostly seen within intact fibers in the normal samples (Fig. 2A), but some normal samples (Fig. 2B) and most painful samples (Fig. 2C) also contained isolated fibers that appeared fragmented. These fragmented fibers were generally intermixed among intact ones located within axon bundles. One painful sample contained only fragmented axons in the coronal pulp (Figs. 1E,F) and these appeared very different than those isolated fragmented axons seen in some normal (Fig. 2B) and most painful samples (Fig. 2C). The expression of Na_v1.7 was minimal in fibers with a fragmented appearance (Figs. 1E,F and 2C), while the expression within fibers with an intact appearance was seen in different-sized fibers and included even staining of low intensity and focal accumulations with brighter intensity. Both of these staining patterns appeared more common in painful samples (compare Figs. 2A to 2C). Three of the painful samples contained localized areas within the middle of the coronal pulp with a high density of fibers with prominent Na_v1.7 staining, but with a diminished intensity or even lack of N52/PGP9.5 staining within these same fibers (Fig. 2D). The Na_v1.7 expression within fibers that lacked N52/PGP9.5 staining would not be evaluated with the quantitative analysis described below. Since increased numbers of accumulations with Na_v1.7 were seen in painful samples, adjacent sections of four sample pairs were stained with caspr, a paranodal protein used to identify nodes of Ranvier 7 and myelin basic protein (MBP) antibodies for further characterization.

Quantitative analysis showed no difference in the Na_v1.7 expression within N52/PGP9.5-identified nerve area or in the pixel intensity of this expression within axons in the pulp horn of normal samples when compared to painful samples (area – 57.6% ± 5.6 vs. 46.6% ± 11.9; Fig. 3A/intensity – 549.6 ± 26.1 vs. 670.8 ± 204.7; Fig. 3B). In contrast, the nerve area occupied by Na_v1.7-immunoreactivity was significantly greater within axon bundles located in the coronal (41.7% ± 4.8 vs. 66.5% ± 5.4; p < 0.01 – Fig. 3C) and the upper radicular (27.2% ± 5.5 vs. 66.0% ± 7.0; p < 0.001 – Fig. 3E) regions of the painful samples, but with no differences in pixel intensity in either of these areas (coronal – 503.8 ± 28.2 vs. 600.2 ± 65.4; Fig. 3D/upper radicular – 489.1± 63.4 vs. 564.8 ± 38.2; Fig. 3F).

Evaluation of pulpal tissue sections stained with caspr allowed a characterization of Na_v1.7 at caspr-identified nodal sites that were classified as either typical or atypical as based on caspr relationships (see Methods for details). This analysis was limited to axon bundles within the coronal and upper radicular regions of four normal and four painful samples, since nodes are commonly seen in these regions. This analysis was performed on images that were thresholded to eliminate low intensity pixels while maintaining those with higher immunofluorescence intensity, since these are the ones typically seen at nodes of Ranvier (Figs. 4A–D). The total number of nodal sites evaluated was 2320 and included more than 500 sites in each region in both normal and painful samples. This analysis showed a dramatic increase in the percentage of nodal sites with Na_v1.7 expression in painful samples within coronal and upper radicular regions (Figs. 4E,F). This increased percentage of Na_v1.7-positive nodal accumulations within painful samples was seen at both typical (coronal – 2.7% ± 1.5 vs. 35.3% ± 5.5; p < 0.01 – Fig. 4E/upper radicular – 5.9% ± 2.7 vs. 32.7% ± 9.8; p < 0.05 – Fig. 4F) and atypical (coronal – 6.9% ± 4.0 vs. 53.9% ± 11.1; p < 0.01 – Fig. 4E/upper radicular – 9.1% ± 5.5 vs. 59.8% ± 8.4; p < 0.01 – Fig. 4F) nodal sites. In addition, there was a significant increase in the occurrence of atypical nodal forms within painful samples relative to the total number of all caspr-identified sites (coronal – 11.0% ± 3.9 vs. 36.9% ± 7.7; p < 0.05 – Fig. 4G/upper radicular – 9.9% ± 3.7 vs. 24.9% ± 4.6; p < 0.05 – Fig. 4H). Together, these results identify the increased association of Na_v1.7 at both typical and atypical nodal sites and an increased occurrence of atypical nodal forms within painful samples.

Pulpal tissue sections that were triple-stained with Na_v1.7, caspr and MBP antibodies allowed a characterization of Na_v1.7 expression at typical and atypical nodal sites in relation to state of myelination as reflected with the use of the MBP antibody (Fig. 5). The expression of Na_v1.7 in normal samples was seen evenly distributed within small fibers that lacked caspr and MBP staining and that most likely represent unmyelinated fibers, while expression within myelinated fibers with caspr and MBP stainings was mostly confined to the nodal site of some fibers (Fig. 5A). The expression of MBP within normal samples was prominent on the surface of fibers located within axon bundles and less at nodal regions (Fig. 5A). In contrast, this expression of MBP was altered in painful samples (Fig. 5B,C). Alterations in MBP included generalized (Fig. 5B) and focal (Fig. 5C) decreases in expressions. Generalized loss was seen in some fibers with typical nodes (Fig. 5B), while localized alterations included a loss of MBP in the area adjacent to heminodes (Fig. 5B) and the axonal region between split nodes (Fig. 5C). Typical and atypical nodes that were associated with a generalized loss of MBP staining (Fig. 5B) and the axonal area located between the split nodes (Fig. 5C), both contained prominent Na_v1.7-immunofluorescence. These findings are consistent with a remodeling of NaChs in areas of demyelination as identified by a loss of MBP staining.

This study was approved by the Human Subjects Institutional Review Board at the University of Texas Health Science Center at San Antonio and informed consent was obtained from all subjects who participated in the study. Teeth were obtained from subjects having an extraction of a normal, nonpainful third molar with fully formed apices (n = 13; 10 females and 3 males with an age range of 19–47), or a painful molar tooth diagnosed with irreversible pulpitis (n = 13; 10 females and 3 males with an age range of 24–49). Painful samples were limited to those that were associated with self-reports of moderate-to-severe levels of pain severity and the presence of spontaneous pain episodes as rated over the 24 hour time-period preceding the extraction. All painful teeth had the presence of a carious lesion that extended into the pulpal tissues, whereas those with large necrotic pulpal lesions were excluded.

Extracted teeth were collected in 0.1 M phosphate buffer (PB) and stored at 4°C. Later the same day, the teeth were split longitudinally and the pulpal tissues were removed and fixed in 4% paraformaldehyde in 0.1 M PB for 20 minutes. The pulpal tissue was rinsed in 0.1 M PB and then placed in 30% sucrose in 0.1 M PB overnight at 4°C. The next day the pulp was placed in Neg-50 (Richard-Allan Scientific; Kalamazoo, MI) and stored at -80°C. Pulpal samples were thawed and embedded in Neg-50, with a normal sample next to a painful sample, and serially sectioned with a cryostat at 30 μms in the longitudinal plane. Sections were placed onto Superfrost Plus slides (Fisher Scientific, Pittsburgh, PA), air dried and then stored at -20°C.

Primary antibodies: rabbit polyclonal anti-Na_v1.7 13 (produced against a 15 amino acid sequence in the rat that shows 13/15 similarities with the human sequence, 1:100 for caspr analysis and 1:200 for nerve area analysis), mouse monoclonal anti-neurofilament 200 kD (N52, Sigma-Aldrich, St. Louis, MO, catalog #N0142, 1:2000), guinea-pig polyclonal anti-protein gene product 9.5 (PGP9.5, Chemicon, Temecula, CA, catalog # AB5898, 1:300), mouse monoclonal anti-caspr (developed by and/or obtained from the UC Davis/NINDS/NIMH Neuromab Facility, supported by NIH grant U24NS050606 and maintained by the Department of Pharmacology, School of Medicine, University of California, Davis, CA 95616, catalog #75-001, 1:500), rat monoclonal anti-myelin basic protein (MBP; Chemicon, Temecula, CA, catalog #MAB386 1:200). Secondary antibodies: Alexa Fluor^® 568 goat anti-rabbit IgG, Alexa Fluor^® 488 goat anti-mouse IgG, Alexa Fluor^® 633 goat anti-rat IgG and goat anti-guinea-pig IgG (all from Molecular Probes, Eugene, OR). All secondary antibodies were used at 1:100.

Immunostaining was performed as described previously 44. Briefly, tissue sections were permeabilized and blocked for non-specific protein binding sites with blocking solution consisting of 4% normal goat serum (Sigma), 2% bovine gamma-globulin (Sigma), and 0.3% Triton X-100 (Fisher Scientific) in PBS for 90 minutes prior to the incubation with primary antibodies in blocking solution for 16 hours. One section from each normal/painful pair was stained with Na_v1.7/N52/PGP9.5, while the adjacent section from select sample pairs were stained with Na_v1.7/caspr/MBP. Sections were rinsed with PBS, incubated in secondary antibody in blocking solution for 90 minutes, rinsed in PBS and ddH₂O, dried, and coverslipped with Vectashield (Vector Laboratories, Burlingame, CA). All procedures were performed at room temperature.

Sections were evaluated with a Nikon Eclipse 90i C1si laser scanning confocal microscope with a 40×/1.30 N.A. oil immersion objective. This evaluation included the selection of a laser gain setting used to image Na_v1.7-immunofluorescence to avoid saturated pixels so to allow a full dynamic range of Na_v1.7-immunofluorescence pixel intensity and to select laser gain settings used to image all other primary antibodies. A series of optical images at 6 μm increments along the "z" axis of the sections stained with Na_v1.7/N52/PGP9.5 antibodies were acquired from the middle 18 μm of each 30 μm thick section in the pulp horn and axon bundles in the coronal and radicular regions when present. Images obtained from the pulp horn in painful samples included only those areas where inflammatory cells were seen nearby to fibers. Care was taken to select fiber bundles that showed N52/PGP9.5 staining within intact fibers, since extensively fragmented fibers may represent degenerating ones. These selection criteria allowed the collection of a z-series of optical images from the pulp horn in 8 normal and 6 painful samples, coronal axon bundles in 12 normal and 8 painful samples, and radicular axon bundles in 8 normal and 8 painful samples. A z-series of optical images were also obtained at 1 μm increments from the middle 20 μm of the 30 μm thick sections of coronal and radicular axon bundles in four normal and four painful samples stained with Na_v1.7/caspr/MBP antibodies. All images were captured with identical settings (including laser gain settings) at a 1024 × 1024 resolution (where each pixel is ≈ 0.3 μm width with an area ≈ 0.09 μm²) and saved as 12 bit single-channel images with a 0–4095 range of pixel intensities. Images were processed for illustration purposes by using Adobe Photoshop CS2 (Adobe Systems, San Jose, CA) and CorelDRAW 12 (Corel Corporation, Ottawa, Canada). Control tissue specimens were processed as above except the undiluted Na_v1.7 antibody was preincubated with peptide antigen (approximately 30:1 peptide to antibody molar concentration ratio) for a minimum of 4 hours before application of the peptide-blocked antibody. Additional control specimens lacked the application of primary antibodies. Control sections lacked specific immunofluorescence.

The specimens stained with the Na_v1.7/N52/PGP9.5 antibodies were used to evaluate Na_v1.7 expression within N52 and PGP9.5 identified nerve fibers. Each individual channel was opened with NIH ImageJ software 45 with the use of the ICS Opener plug-in 46. The corresponding PGP9.5 and N-52 z-slices were combined by using the "Max Operation" in the Image Calculator. The staining intensities in all images were filtered by applying a threshold value to remove low intensity pixels that represent nonspecific/background values. The threshold value was determined with mean and standard deviation pixel intensity values obtained from a histogram analysis of every Na_v1.7 and combined N52/PGP9.5 image slice from normal samples. Average values for mean and standard deviation pixel intensities were then obtained independently for the Na_v1.7 and combined N52/PGP9.5 groups, and then a threshold value (mean + 2 times the standard deviation) for each group was applied to remove (filter) background staining. This thresholding process was applied in a consistent manner to all images from the normal and painful sample groups and resulted in images in which N52/PGP9.5 staining was clearly limited to nerve fibers. The thresholded nerve fiber area was recorded by creating a selection in every slice and this was saved as a region of interest (ROI). The nerve area ROI was transposed onto the corresponding thresholded Na_v1.7 image. The thresholded image was then redirected to the original Na_v1.7 image so that when it was analyzed, the actual Na_v1.7-immunofluorescence pixel intensities within the ROI were recorded. The particles were then analyzed, and the average intensity and percent area of nerve fiber occupied by Na_v1.7 were recorded.

The Na_v1.7 expression within single fibers at caspr-identified nodal sites that were classified as either typical or atypical were determined in the z-series of images that were obtained in sections stained with the Na_v1.7 and caspr antibodies. The typical nodal sites were identified by extensive caspr staining within two bands that were separated by a nodal gap, while any deviations from this relationship were classified as atypical nodal sites. The most common caspr-identified atypical nodal site included heminodes that were identified as a single band of caspr staining, or two bands of caspr staining separated by a nodal gap but where the area of caspr staining was greatly diminished on one side to less than 50% of that seen on the side with more extensive caspr staining. All caspr-identified sites within single fibers were classified in each z-series image stack with navigation through each stack to clarify relationships, while the Na_v1.7 expression at individual nodal sites was performed on maximum intensity collapsed z-projection images generated from each z-series. The Na_v1.7-immunofluorescence pixel intensity in each collapsed image was filtered by applying a threshold value to remove low intensity pixels while leaving pixels with higher immunofluorescence intensity like those located within nodal accumulations. A threshold value for Na_v1.7 staining was determined with mean and standard deviation pixel intensity values obtained from a histogram analysis of the single-channel Na_v1.7-only maximum projection from each of the four normal tooth pulp samples. Mean values were calculated, and a threshold value (mean + six times the standard deviation) was applied to each maximum intensity collapsed z-projection image. All caspr-identified nodal sites were included in the analysis and Na_v1.7-positive nodal sites were defined as those accumulations with one or more Na_v1.7-positive pixel(s) above threshold that were clearly associated with caspr staining as confirmed with identification of the same nodal site in the appropriate z-series image stack. This analysis was limited to caspr-identified sites located within fibers that were fully represented within the z-series image stack. All nodal accumulations were classified as either typical, or atypical and further classified as associated with Na_v1.7 or as lacking this association. This evaluation further allowed the characterization of Na_v1.7 at some caspr-identified atypical sites that were classified as split nodes where two distinct Na_v1.7 accumulations were separated by a gap in the Na_v1.7 staining within the same fiber and with each Na_v1.7 accumulation being flanked on only one side by caspr.

Statistical analysis to determine significance used the unpaired Student t-test while error bars on all graphs represent the standard error of the mean (SEM). The "n" values used in all instances represented the number of samples evaluated.