Anti-high mobility group box 1 antibody suppresses local inflammatory reaction and facilitates olfactory nerve recovery following injury

This study was performed using transgenic mice (OMP-tau-lacZ mice) obtained from the Jackson Laboratory (Bar Harbor, ME, USA), whose strain is derived from C57BL/6 mice. In this strain, the gene sequence encoding the olfactory marker protein (OMP) has been replaced with a tau-lacZ reporter gene [26]. The OMP is expressed in all mature olfactory neurons [27], and the replacement with tau-lacZ reporter gene enables the visualization of olfactory nerve fibers and their projections to olfactory bulb glomeruli. The advantage of using these mice is that a histological assessment of degenerating and regenerating olfactory nerve fibers can be performed using a standard method for staining and light microscopy. Although the exact function of OMP and effects of lack of OMP has not been revealed yet, many previous studies have shown that OMP-tau-lacZ mice are capable of recovering olfactory function after olfactory nerve injury [13, 14].

Both male and female adult mice were used in this study and randomly assigned to experimental groups. Mice were anesthetized with sodium pentobarbital (80 mg/kg, ip). Under sufficient anesthesia, a frontal craniotomy was performed to expose the olfactory bulbs. An olfactory nerve transection procedure (NTx) was performed between the olfactory bulb and cribriform plate using a curved rigid stainless steel blade to generate a severe olfactory nerve injury model [15]. For histological assessments, the NTx procedure was performed only on the left side (injury side) of the animal while the right side (right olfactory bulb and nerves) remained intact and served as an internal histological control (Fig. 1). For other assessments (PCR, olfactory function testing, evoked potentials), a bilateral NTx was performed, cutting the olfactory nerves to the right and left olfactory bulbs, resulting in a complete loss of smell (anosmia). After the NTx procedure was complete, the skin incision was sutured and the animal closely monitored until it was awake and fully recovered from anesthesia. All protocols and surgical procedures for this study were reviewed and approved by the Institutional Animal Care and Use Committee of Mie University.

The olfactory bulb tissue samples were placed into Sepasol Super G and homogenized. Total RNA was extracted and purified by DNase treatment according to the manufacturer’s protocols. The reverse transcription reaction was performed using the ReverTraAce (Toyobo, Japan) using 1 μg of total RNA. Variation in the reverse transcription reaction was limited by transcribing all samples simultaneously. PCR reactions were optimized to 94 °C for 2 min, 35 amplification cycles for HMGB1 and 26 cycles for GAPDH at 94 °C for 10 s, 60 °C for 20 s, 72 °C for 40 s, and a final extension of 2 min at 72 °C. Amplified products were resolved on 1.5% agarose gel and visualized by ethidium bromide staining using ITLS software.

Primer Sequences

HMGB1: Primer Bank ID 6754208a1

Forward: GGCGAGCATCCTGGCTTATC

Reverse: GGCTGCTTGTCATCTGCTG

Amplicon size: 86 bp

GAPDH: Primer Bank ID: 126012538c3

Forward: TGGCC TTCCG TGTTC CTAC

Reverse: GAGTT GCTGT TGAAG TCGCA

Amplicon size: 178 bp

To investigate the associations among HMGB1, the inflammatory reaction, and nerve regeneration after injury, the anti-HMGB1 antibody (Novus Biologicals, USA) was injected intraperitoneally just after the NTx. To determine if there is a dose-dependent effect of the drug, low (25 μg in 0.5 ml saline) and high (50 μg in 0.5 ml saline) doses of the anti-HMGB1 antibody were used, making them comparable to a dose used in a previous study [25]. We also injected IgG (50 μg in 0.5 ml saline) intraperitoneally in a group of control animals. We collected data from six mice for each of the three treatment groups and each of the four recovery time points (days 5, 14, 42, and 100) for a total of 72 mice (6 mice × 3 doses × 4 recovery points).

Tissue sections were washed at room temperature with buffer A [100 mM phosphate buffer (pH 7.4), 2 mM MgCl₂, and 5 mM EGTA (ethylene glycol tetraacetic acid)] once for 5 min and then a second time for 25 min. This was followed by two 5-min washes with buffer B [100 mM phosphate buffer (pH 7.4), 2 mM MgCl₂, 0.01% sodium deoxycholate, and 0.02% Nonidet P40]. The blue X-gal reaction was generated overnight in the dark by exposure to buffer C (buffer B, with 5 mM potassium ferricyanide, 5 mM potassium ferrocyanide, and 1 mg/ml of X-gal). The X-gal reaction was stopped by two 5-min washes in phosphate buffer.

After confirming the appearance of the blue X-gal reaction, tissue sections were counterstained with a 1% Neutral Red solution. Sections were examined and digitized using CCD photomicroscopy. Areas of injury-associated tissue, including inflammatory cells and glial scar tissue, were identified along with blue (X-gal) labeled olfactory nerve endings within the glomerular layer of the olfactory bulb (Fig. 1). The area of injury-associated tissue was outlined on digital images of tissue sections and quantified using ImageJ (ver.1.50i, National Institute of Health [NIH], USA) software. For the measurement of injury-associated tissue, we targeted an area that was surrounded by the following four margins: the cribriform plate as an anterior margin, anterior edge of the olfactory bulb as a posterior margin, a line connecting the posterior end of the nasal septal mucosa and the anteriomedial corner of the olfactory bulb as a medial margin, and a line connecting the posteriolateral end of the most lateral sinus and the anteriolateral corner of the olfactory bulb as a lateral margin. Since this targeted area is where the olfactory nerve fibers normally run from the sinonasal mucosa to the olfactory bulb before NTx, as seen in the control side, the tissue levels in this area can be associated with the degree of nerve degeneration and regeneration.

The area (mm²) of tissue observed between the cribriform plate and olfactory bulb (Fig. 1) was measured in two representative horizontal sections (sections A and B) from each animal and averaged. Section A was selected to represent the dorsal level. At this particular level, a large olfactory nerve bundle is observed passing from endoturbinate II through the cribriform plate to the olfactory bulb (Fig. 1). Section B represented a more ventral level. At this level, endoturbinate III attaches to the cribriform plate. The area measurements from NTx mice at each of the four recovery time points were used to compare mean values for injury-associated tissue. The levels of olfactory nerve degeneration and regeneration were assessed by comparing changes in the amount of blue X-gal staining in the glomerular layer on the left (NTx injury side) to that on the right (control) side. Horizontal olfactory bulb sections (sections A and B) were also used to obtain measurements of (1) the glomerular layer perimeter distance (G-P distance), a continuous line passing through the center of all the glomeruli within the bulb section, and (2) the total length of glomerular segments along the perimeter that were labeled with the blue X-gal stain (G-X-gal distance). The ratio of the X-gal-stained distance (G-X-gal distance) to the total perimeter of the glomerular layer (G-P distance) was obtained for both the NTx injury and control sides. Changes in the blue X-gal nerve staining on the NTx injury-left side were expressed as percentage of the X-gal staining on the intact control side and were used to measure levels of olfactory nerve degeneration and regeneration within the olfactory bulb, as follows:

Immunohistochemical staining for HMGB1 protein, glial fibrillary acidic protein (GFAP), and cluster of differentiation (CD) 68 glycoprotein was performed on horizontal sections of NTx or control mice. HMGB1 staining was performed at two different time points following left NTx injury, 12- and 24-h points. GFAP and CD68 staining was performed at four different time points following left NTx injury, days 5, 14, 42, and 100. GFAP is constitutively produced by astrocytes. In the reactive glial response to the central nervous system injury, hypertrophic reactive astrocytes increase their expression of GFAP [28]. CD68 staining was used to measure injury-induced inflammatory changes at different time points after NTx injury. CD68 is a lysosomal membrane-associated glycoprotein that is expressed on the surface of histiocytes, cells that are part of the immune system, including macrophages and microglia and play an important role in phagocytic activities.

After washing with phosphate-buffer saline (PBS) for 5 min, sections were processed by immersion for 1-min intervals in a series of alcohol solutions (70, 95, 100, 95, and 70% ethanol). This was followed by three 5-min washes with 0.3% Triton X-100 in PBS. Sections were then incubated with 5% normal goat serum, 1% bovine serum albumin, and 0.5% Triton X-100 in PBS for 30 min and reacted with one of the following primary antibodies: rabbit anti-HMGB1 antibody (Novus Biologicals, USA), rabbit anti-mouse GFAP antibody (1:500, DAKO, USA), and rat anti-mouse CD68 antibody (1:100, AbD serotec, USA). Anti-HMGB1 antibody was visualized using biotinylated goat anti-rabbit IgG (Vector Laboratories, USA) and horseradish peroxidase streptavidin (Vector Laboratories, USA). Diaminobenzidine (DAB) was used as a chromogen. Anti-GFAP and CD68 antibodies were visualized using Cy3-conjugated goat anti-rabbit IgG (1:100, GE, USA) and Alexa Fluor 488-conjugated goat anti-rat IgG (1:100, Invitrogen, USA) under fluorescent microscope, respectively. GFAP- and CD68-positive cells were counted in five different 0.01-mm² sampling areas located in the anterior (injured) region of the olfactory bulb (five areas: the anterior apex area, areas of anteriomedial corner and anteriolateral corner of the olfactory bulb, and fixed midpoint areas between the anterior apex area and anteriomedial and anteriolateral corner areas). The average number of GFAP- and CD68-positive cells/0.01mm² was then calculated for NTx mice at each of the four recovery time points.

To determine if olfactory function recovered after the NTx, a smell detection test using avoidance conditioning behavior to cycloheximide was administered to mice before and after the NTx as reported previously [13, 14]. Cycloheximide has a peculiar odor and unpleasant taste for mice. Normal mice were first deprived of water for 48 h and then trained to avoid cycloheximide solution. Before NTx surgery, mice were conditioned in two or more training sessions, each consisting of 10 trials. In each trial, the mouse was presented with bottles of 0.01% cycloheximide solution and distilled water, one positioned on the left, the other on the right side of a test cage. When the mouse licked the delivery tube of either bottle, the bottles were withdrawn from view and presented again. The left and right positions of the two bottles were shifted according to the Gellermann series (cycloheximide bottle position: right (R)-left (L)-L-R-L-L-R-R-R-L). Mice were considered to have learned the smell of cycloheximide when they chose the distilled water bottle 10 consecutive times out of 10 trials (percent score: 100%) on two consecutive test sessions. After NTx surgery, the test was administered every 7 days until the mouse regained its olfactory function (scored 10 out of 10 correct responses) or exceeded a 100-day cutoff period. Mice that scored 100% at one of the recovery test days were considered to have fully recovered their olfactory function.

To confirm that regenerated olfactory receptor cell axons were functionally reconnected to the olfactory bulb after the olfactory NTx, field potentials evoked by electrical stimulation of the olfactory mucosa were measured in the olfactory bulb of mice that completed the cycloheximide olfactory function test as reported previously [13, 14]. For this group of mice, the NTx procedure was bilateral, resulting in a complete loss of olfactory function. On day 100, the animal was anesthetized and the nasal bone removed to expose the olfactory mucosa, and a second craniotomy was performed to expose the olfactory bulb. The olfactory mucosa was electrically stimulated using a concentric circular needle electrode (200 μm in diameter) that delivered a constant current stimulus of 0.5 mA, 0.3 ms, at a rate of 1 Hz. Field potentials were recorded from a region of the ipsilateral olfactory bulb 1000 μm lateral to the midline, 6000 μm rostral to the bregma, and 1000 μm ventral to the surface. The recording microelectrodes used in this experiment were glass capillary pipettes filled with 0.5 M KCl with a tip resistance of 3–6 MΩ. To improve the signal-to-noise ratio of recordings, the olfactory mucosa stimulation-induced responses were averaged for 32 stimulus trials. The magnitude of field potential was quantified by measuring the time integral of the evoked field potential.

All numerical data obtained are expressed as means ± standard error of the mean (SEM). For statistical analysis of the data, the Mann–Whitney U test was used to determine differences in average values between two groups. For three groups, the two-way ANOVA was used, and post hoc comparisons were performed by the Fisher’s PLSD method for a gene expression study and by the Bonferroni’s method for analyses of a histological study. The chi-square (χ²) test for independence was used to test for differences in ratio. Differences were regarded as significant when p < 0.05 for two group and p < 0.016 for three group comparisons.

A semi-quantitative PCR shows that there are significant increases in HMGB1 gene expression 12 h after the NTx or later, with a peak 12 h after the NTx injury before beginning to decrease (Fig. 2). Immunohistochemical staining using anti-HMGB1 antibody showed strong positive reaction in the injury-associated tissue and damaged anterior part of the olfactory bulb of the NTx side compared to the control side (Fig. 3). These results showed that HMGB1, the target of its antibody, increased at damaged olfactory tissues after the NTx.

To determine if anti-HMGB1 antibody treatment can facilitate recovery of the olfactory nerves after NTx injury, it was injected intraperitoneally in the severe injury model. Figure 4a shows results of the control (IgG) compared to effects of anti-HMGB1 antibody treatment (Fig. 4b). A decrease in the percentage of X-gal (blue) staining on the NTx side at day 5 and day 14 reflects the degeneration of olfactory nerves (Fig. 4c). However, the subsequent increase in blue staining in the nerve and glomerular layers at days 42 and 100 indicate that the regenerating olfactory nerves had reestablished connections with the olfactory bulb. Compared to the IgG controls, a significantly higher level of the nerve recovery was found in the anti-HMGB1 antibody-injected mice at day 100, and this increase was dose-dependent.

Figure 4d shows changes in the amount of injury-associated tissue (glial scar) present on the NTx side. The amounts increased at day 5 and gradually decreased at day 14 and after. The tissue amount in low and high doses of anti-HMGB1 antibody-injected mice was significantly less than that in the control IgG mice.

Both GFAP-positive cells and CD68-positive cells increased on the NTx side in the olfactory bulbs at day 5 and gradually decreased at day 14 and later recovery times (Fig. 5). With anti-HMGB1 antibody treatment, the number of both GFAP and CD68 cells decreased compared with those in control IgG mice in a dose-dependent manner.

An olfactory function test using avoidance conditioning behavior to cycloheximide was administered to mice injected with anti-HMGB1 antibody (50 μg in 0.5 ml saline) and control IgG (50 μg in 0.5 ml saline) before and after the NTx. For the anti-HMGB1 antibody group, 11 of 16 (69%) mice achieved a score of 100% on the olfactory function test after NTx, indicating that their olfactory function had recovered (Table 1). The average time required for behavioral recovery in the 11 mice was 48 ± 7 days. For the control saline group, however, only 3 of 15 (20%) mice recovered their olfactory function (98, 77, and 35 days, 70 ± 19 days). Anti-HMGB1 antibody-injected mice showed a significantly higher rate of olfactory function recovery than control IgG mice (p < 0.05).

Positive field potentials induced by electrical stimulation of olfactory mucosa were successfully recorded from the mice that showed functional recovery in the olfactory test while little or no field potentials were observed in mice that did not recover olfaction function (Fig. 6). The anti-HMGB1 antibody group (n = 16) showed significantly larger field potentials than the IgG control group (n = 15).