Eighty-three Sprague-Dawley rats (200–250 g) were randomly assigned to the control, sham-operation, and NPexposed groups (Fig. 1). For the analysis of AMPA receptor activity, the rats (n=40) were assigned to the control (n=6), sham-operation (n=4), and NP-exposed (n=30) groups. For assessing the immunohistochemical expression of OX-42 and glial fibrillary acidic protein (GFAP), the rats (n=40) were assigned to the control (n=10) and NP-exposed (n=30) groups. Moreover, three rats were used to investigate the progression of neuropathic pain. They were housed, two per cage, with diurnal light-dark cycles, and food pellets and water were provided ad libitum.

All experimental procedures were conducted according to the guidelines of the Institutional Animal Care and Use Committee at Yeungnam University, Republic of Korea (No. YUMC-2017-010). The rats were anesthetized through the intraperitoneal injection of Zoletil 50 mg/kg (Virbac, Carros, France). With the animals placed in the prone position, a midline dorsal incision was made within the pelvic area, and the multifidus muscles were separated along the L4–S1 spinous processes. Laminectomy was performed to expose the left L5 nerve roots and DRG. An incision, approximately 1 cm in length, between the second and third coccygeal vertebrae of each tail was made on the dorsal surface of the proximal tail for harvesting the autologous NP (Fig. 2). The NP was harvested, using a curette, and implanted beside the left L5 nerve root, immediately proximal to the DRG. The sham surgery was performed similarly, but without the implantation of autologous NP in the left L5 nerve root.

Mechanical allodynia was determined by measuring the paw withdrawal response to the mechanical stimulation of the ipsilateral hind paws with von Frey filaments (North Coast Medical Inc., Morgan Hill, CA, USA); it was calibrated in grams. The rats were placed in a clear plastic cage with a metal mesh floor and permitted to adapt to the testing environment for 15 minutes, and the plantar surface of each hind paw was stimulated to cause a slight bending of the von Frey filaments for 3–5 seconds. The filaments were stimulated in the order of increasing and decreasing thicknesses, beginning with a 0.1 g probe, and the stimulation was maintained until a consistent withdrawal response to >3 out of 5 stimuli was observed. If no withdrawal response was elicited by the 15 g filament, the mechanical threshold was set at 15 g. The mechanical allodynia at the plantar surfaces of the ipsilateral hind paws was assessed three times a week until the animals were sacrificed. Three rats were tested for up to 38 days postoperatively to identify the pattern of progression of neuropathic pain (Fig. 3).

Five rats each were euthanized 3, 7, and 14 days after surgery for the immunohistochemical staining of the spinal cord for the GluA1 and GluA2 subunits of the AMPA receptor OX-42 and GFAP. The rats in the control group were euthanized without performing sham surgery. Under anesthesia, a catheter was inserted from the left ventricle of the heart to the aorta, and the animals were perfused with heparinized saline followed by perfusion with 500 mL of 4% paraformaldehyde in phosphate buffer (0.1 M, pH 7.4). The L5 spinal cord segment was excised and fixed for 24 hours with the perfused fixative, and the tissue samples were cryoprotected in 30% sucrose made in phosphate buffer. Transverse sections of the spinal cord (30-μm thick) were sliced on a cryostat and blocked with 10% normal donkey serum (Jackson ImmunoResearch Laboratories Inc., West Grove, PA, USA) in 0.01 M phosphate-buffered saline (PBS) containing 1% Triton X-100 for 40 minutes to decrease the non-specific binding of the primary antibody. For performing immunofluorescence microscopy, the sections were incubated overnight at room temperature with a mixture of AntiGluR1 (ABN241, dilution 1:1000; Millipore, Burlington, MA, USA) and anti-GluR2 (AB1768-I, dilution 1:2000; Millipore) antibodies, or with mouse anti-OX-42 (dilution 1:200; Chemicon, Temecula, CA, USA) and anti-GFAP (dilution 1:1000; Dako, Santa Clara, CA, USA) antibodies. After primary antibody incubation, the sections were rinsed with PBS 3 times (5 min/rinse) and incubated with 2% normal donkey serum for 15 minutes followed by an appropriate combination of Alexa 488-conjugated donkey anti-rabbit secondary antibodies (dilution 1:200; Invitrogen, Eugene, OR, USA) containing 1:200 dilution of Triton X-100 for 2 hours. After several rinses postincubation, the sections were mounted on coverslips covered with a VECTASHIELD Mounting Medium (Vector Laboratories, Burlingame, CA, USA) and observed under a fluorescence microscope (Digital Module R, Leica Microsystems, Switzerland). Images were visualized and captured by a charge-coupled device camera (Olympus, Tokyo, Japan) attached to the microscope.

To analyze the immunopositivity of AMPA receptors and glial cells in the spinal dorsal horn, images obtained from the L5 spinal cord sections were captured. One image (898×660 μm) captured by the camera was encoded to blind the investigator before analysis. The number of pixels positive for AMPA receptors and glial cells was segmented by applying an appropriate threshold. The intensity of green fluorescence for each frame was obtained using an image analysis software (Leica application suite V4.2; Leica Microsystems, Switzerland). The superficial dorsal horn of the spinal cord had a connection with nociceptive neurons [15]. AMPA receptors are more in the superficial dorsal horn of the spinal cord [16]. AMPA positive and glial cell activity in the superficial dorsal horn were evaluated [17,18]. The mean values of each group and the time point versus control mean value in the dorsal horn of the L5 spinal level were calculated as percentages. The numbers of OX-42-positive and GFAP-positive cells were counted using the image analysis software.

Data from each rat were pooled and processed using Excel version 16.26 (Microsoft Corporation, Redmond, WA, USA). The collected data were uploaded into SPSS software (IBM Corporation, Armonk, NY, USA) and analyzed. For comparison, the control values were considered as 100. Comparisons of the findings from the quantitative image analysis of immunoreactivities and the measurements of mechanical withdrawal thresholds were performed using the Kruskal-Wallis test. Statistical significance was set at p<0.05.

The mechanical withdrawal threshold of the control and sham-operated groups was 15 g. Rats that underwent sham surgery did not exhibit a decreased mechanical withdrawal threshold 3 days after surgery. Rats in the NP-exposed group that underwent lumbar disc herniation surgery showed statistically significant reductions in the mechanical withdrawal thresholds of the ipsilateral hind paw 3 days after surgery, and these were maintained until postoperative day 31 (Fig. 2).

The immunoreactivities of GluA1 and GluA2 were well-visualized in the spinal dorsal horn, and they showed intense staining in the superficial lamina of the dorsal horn, the region that received nociceptive input from the periphery (Fig. 4). Compared with the control group, the immunoreactivity of GluA1 significantly increased in the animals that underwent surgery within postoperative days 3 to 14. Three and 7 days after the surgery, the increase in immunoreactivity of GluA1 in the spinal dorsal horn was statistically significant (Fig. 5). The immunoreactivity of GluA2 also increased within postoperative days 3 to 7; however, the magnitude of increase was relatively small compared with the immunoreactivity of GluA1. The differences between the immunoreactivities of GluA1 and GluA2 at each time point were not statistically significant. Fourteen days after surgery, the increased immunoreactivities of the GluA1 and GluA2 subunits observed in the rats that underwent surgery gradually normalized to levels observed in homeostatic conditions similar to those of the control group.

Immunohistochemical examination of the spinal dorsal horn for OX-42 and GFAP demonstrated reactive activation of the microglia and astrocytes after surgery (Fig. 6). The immunoreactivity of OX-42 within the superficial lamina of the spinal dorsal horn had significantly increased on postoperative day 7. The immunoreactivity of GFAP also significantly increased on postoperative days 3 and 7 (Fig. 5). On postoperative day 14, the numbers of OX-42-positive and GFAP-positive cells decreased to levels observed in the control group.