Effect of diminished flow in rabbit lumbar arteries on intervertebral disc matrix changes using MRI T2-mapping and histology

This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol of this study was approved by the Institutional Animal Care and Use Committee of Mie University (reference number: 21–11). Twenty-two 12-week-old New Zealand White rabbits (female) weighing 2.4–3.0 kg (Kitayama Labs Laboratory Animals Bleeding & Equipment Supply, Ina, Japan) were housed in separate cages under standard conditions with a light-dark cycle (12 h – 12 h) and dry-bulb room temperature at 22–24 (°C), and provided ad libitum access to tap water and food pellets daily.

Following intramuscular administration of 25 mg/kg ketamine hydrochloride (Ketalar®, Daiichi Sankyo, Tokyo, Japan) and 5 mg/kg xylazine hydrochloride (Selactar®, Bayer Yakuhin, Tokyo, Japan), lateral plain radiographs were taken to determine preoperative baseline values for IVD height. After sedation and induction, anesthesia was maintained with 3.0% isoflurane (Schering-Plough Animal Health, Tokyo, Japan) using anesthesia mask (Acoma medical, Tokyo, Japan). The rabbits were placed in a lateral prone position, and a posterolateral retroperitoneal approach exposed the abdominal aorta from the left renal artery to the aortic bifurcation. Lumbar arteries L2-L6 were isolated, and the third and fourth lumbar arteries providing blood flow to L4 and L5 vertebrae were doubly ligated using vascular clips (Ligaclip®, Ethicon, Somerville, NJ, USA) (Fig. 1). The disruption of blood flow (collapse of lumbar arteries) just distal to the ligation site was confirmed under direct view through a magnifying glass (× 2.9 magnification, Surgical Acuity™, Middleton, WI, USA) in all the rabbits. Blood flow to the L3/L4 disc (cranial disc) was diminished by ligation of the third lumbar artery, and blood flow to the L5/L6 disc (caudal disc) was diminished by ligation of the fourth lumbar artery. The L4/L5 disc was termed the bilateral disc because blood flow to the L4/L5 disc was diminished by ligation of both the third and fourth lumbar arteries. The L2/L3 disc was used as the control disc (Fig. 2). The wound was closed with layered sutures. The rabbits were returned to their cages after recovery observation.

Among seventeen rabbits that operated, two rabbits showed paralysis of the hind legs immediately after the surgery and not included in the experiment. These two rabbits were euthanized by carbon dioxide inhalation following premedication by the intramuscular administration of ketamine hydrochloride (25 mg/kg). 100% carbon dioxide gas was introduced into the chamber where the rabbits were placed (with its fill rate at approximately 10 to 30% of the chamber volume per minute).

The remaining fifteen rabbits showed no abnormality movement of the hind legs, and showed similar body weight changes throughout the experimental period.

Under general anesthesia induced by the intramuscular administration of ketamine hydrochloride (25 mg/kg) and xylazine hydrochloride (5 mg/kg), lateral plain radiographs of the lumbar spine were taken before surgery and at 2, 4, 6, 8, 10, and 12 weeks after surgery. At 4, 8, and 12 weeks after surgery, five rabbits per group were euthanized as described above. Intact spinal columns were harvested for magnetic resonance imaging (MRI) and histological analyses to determine the degree of degeneration. Five 12-week-old rabbits with a sham operation that only exposed the abdominal aorta were used for a 0-week control group.

Disc height was radiographically monitored biweekly from the day of operation to 12-weeks post-operation (Fig. 3). The soft X-ray radiograms were digitized and measurements of vertebral body height and disc height using OsiriX Imaging Software (OsiriX Foundation, Geneva, Switzerland) were made by an orthopedic researcher blinded to the treatment group. Intervertebral disc height was expressed as the disc height index (DHI), as previously described [18]. A change in the DHI for a disc was expressed as percent disc height index (%DHI) and normalized to the measured preoperative disc height (%DHI = [postoperative DHI/ preoperative DHI] × 100).

After the rabbits were sacrificed, the L1 to L6 vertebrae with surrounding soft tissues were isolated and wrapped by plastic cling film to prevent dehydration, and kept in ice box until subjecting to quantitative transverse relaxation time (T2) MRI analysis as previously reported [19]. MRI was performed using a 3.0-Tesla imager (Achieva 3.0 T, PHILIPS, Amsterdam, The Netherlands) with a 3″ birdcage extremity coil (PHILIPS). Room temperature was kept constant at 24 °C during the MRI scan. T2 mapping was performed using a multi-echo spin-echo sequence in the sagittal plane. Scanning parameters were: Repetition Time (TR) = 3000 ms, Echo time (TE) = 20, 40, 400 ms (20 TEs), field of view = 10 cm, slice thickness = 2 mm, image matrix = 560 × 560, number of excitation = 1. Total scanning time per sample was 10 min 39 s. For creating color-coded T2-maps, MR images at multiple TE were imported into OsiriX Medical Image software by a T2 mapping plug-in.

MRI analysis was performed on IVDs L2/L3, L3/L4, L4/L5 and L5/L6. Mean signal intensities were determined in the regions of interest (ROI) framing the inner border of the AF on T2-weighted images taken at TE of 100 ms (Fig. 4A). Images were selected that best visualized the borders of the AF and NP (Fig. 4B-E).

Intervertebral disc samples were harvested at 4, 8, and 12 weeks after surgery to analyze the degree of degeneration over time. After MRI assessment, the experimental IVDs were fixed in 4% formalin, soaked in a decalcifying agent (K-CX: FALMA, Tokyo, Japan), embedded in paraffin, and sectioned midsagittally. The samples were stained with hematoxylin and eosin for analysis of cellular constituents and Safranin-O for PG content. An observer, blinded to the experiment, analyzed the histological sections and graded them using a previously reported protocol [18, 20]. Histological grading of midsagittal sections was performed based on a scale using four categories of degenerative change, including (A) patterns of fibrocartilage in the lamellae of the AF, (B) the border between the AF and NP, (C) the cellularity of the NP, and (D) the matrix of the NP, as well as (E) the histological changes of the pericellular matrix of NP cells (Table 1). Discs were graded with a value from 1 to 3 in each of the five parameters. A normal disc was assigned a value of 1 and a degenerated disc was assigned a value of 3. Grades ranged from 5 to 15 with a maximum of 3 points for each parameter; therefore, a grade of 15 represents severe degeneration. The total scores for each disc in this semi-quantitative grading assessment were statistically analyzed.

IVD samples (L2/L3 to L5/L6) of five rabbits at 4 weeks after surgery were representatively used for immunohistochemical analysis. Five L2/L3 discs at 4 weeks after surgery were used as the non-ischemic control. Samples were fixed, decalcified, and embedded in paraffin for serial 5 μm sectioning used for immunohistochemical analysis. Following endogenous peroxidase inactivation and heat-induced epitope retrieval, the sections were stained with anti-HIF-1α antibody (H1alpha67: Novus Biologicals, Littleton CO, USA) for immunohistochemical analyses. Mouse IgG (DakoCytomation, Glostrup, Denmark) was used as the isotype or negative control. The sections were visualized using the universal immuno-enzyme polymer method (Histofine Simple Stain MAX-PO; Nichirei Biosciences, Tokyo, Japan) and 3,3′-diaminobenzidine tetrahydrochloride (DAB; Dojindo, Tokyo, Japan), followed by counterstaining with Mayer’s hematoxylin.

The number of immunoreactive cells was counted by one observer who was blinded to the experimental group. The percentage of positive-staining cells in the outer AF (oAF), inner AF (iAF) and NP area was quantified using the mean percentage of immunopositive cells from five fields of microscopic images per sample at 200x magnification.

Differences in %DHI and MRI T2-values were assessed for statistical significance by two-way repeated measures analysis of variance (ANOVA) to compare the IVD levels with time points of examination, followed by the Bonferroni post hoc test. Sample size was determined with a power analysis for an alpha of 0.05 and power of 0.80 using G*POWER3 [21] with a sample size of 5 animals at each time point. Histological grading scores were assessed by the Kruskal-Wallis test for inter-group comparisons and the Friedman test for temporal changes. The correlation between MRI T2-values and histological grading scores was evaluated using Spearman’s rank-order correlation test. The % of immuno-positive cells was evaluated by two-way ANOVA to compare IVD levels and disc area, followed by the Bonferroni post hoc test. All data were expressed as mean ± standard deviation (SD). All the statistical analyses were performed using IBM Statistical Package for Social Sciences Software (SPSS) Statistics (IBM Japan, Tokyo). The accepted level of significance was p < 0.05.

Representative radiographs before and after lumbar artery ligation showed no remarkable narrowing of cranial, bilateral or caudal discs compared to before surgery and the intact L2/L3 disc of the same animal (Fig. 5). The %DHI of each experimental group significantly decreased during the observation period (p < 0.01, 2-way repeated measures ANOVA), although the analysis of variance showed no significant interaction between disc level and time-point (p = 0.83). The %DHI at each disc level changed similarly during the observation period and no significant differences were found among the disc levels (Fig. 5). No significant differences were also seen in the %DHI among the four groups at each time point (Fig. 5).

Representative T2-mapping images of control (L2/L3) discs at 8 weeks postoperative showed that higher T2 values, as indicated by red pixels, were uniformly observed in the NP of IVDs (Fig. 4B). However, the loss of red pixels, replaced by yellow pixels, was identified in the center of the NP of cranial, bilateral, and caudal discs (Fig. 4C-E).

T2-values of the control, cranial, bilateral, and caudal discs (n = 5, respectively at each time point) showed significant decreases during the experimental period (p < 0.01, 2-way repeated measures ANOVA), although the analysis of variance showed no significant interaction between disc level and time-point (p = 0.58). The averaged T2-values of the ischemia group of cranial, bilateral and caudal discs were lower than that of the control L2/L3 group throughout the observation period but failed to reach statistical significance (p = 0.2, 2-way repeated measures ANOVA). No significant differences in T2-value were also found among the ischemia group of cranial, bilateral and caudal discs.

When the data were analyzed at each time point, the average T2-value of the cranial group was significantly lower than that of the L2/L3 control group at 4 weeks postoperative (T2-values [ms]: control: 172.1 ± 10.0, cranial: 159.2 ± 5.6, bilateral: 152.8 ± 7.5, caudal: 146.9 ± 9.5, p < 0.05 vs. control) (Fig. 6). No significant differences were seen in T2-values among the four groups at 0, 8, 12 weeks after surgery (Fig. 6).

Histological changes were observed in NP tissues of cranial, bilateral, and caudal discs (n = 5, respectively at each time point) after lumbar artery ligation (Fig. 7); however no histological changes were found in AF tissues (Fig. 8). At 4 weeks post-ligation, the Safranin-O staining properties of the pericellular matrix of ischemic cranial, bilateral and caudal discs had become irregular (Fig. 7F-H). Circumferential condensation of the ECM, as shown by intense staining for Safranin-O in/around the pericellular matrix, was found in the NP of cranial, bilateral and caudal disc tissues (Fig. 7F-H). At 8 weeks, an irregular and circumferential staining pattern of the pericellular matrix was found in NP tissues of cranial, bilateral and caudal discs similar to those at 4 weeks post-ligation (Fig. 7J-L). At 12 weeks, diffuse staining for Safranin-O was seen in both the pericellular and interstitial matrix in the NP of cranial, bilateral and caudal discs (Fig. 7N-P).

In AF tissues, fibrochondrocyte-like cells were aligned longitudinally between the collagen fiber bundles. However, no remarkable changes in the pattern of fibrocartilage lamellae and serpentine pattern of fibers were found (Fig. 8). No ruptured fibers were identified in the AF tissues.

The total histological grading scores showed that no significant differences among the four groups (n = 5) were seen at 0, 4 and 12 weeks post-surgery (Fig. 9A). However, lumbar artery ligation did significantly affect the total histological score of corresponding IVDs at 8 weeks after the surgery (p < 0.05, Kruskal-Wallis test) (Fig. 9A). The total scores for cranial disc were significantly increased compared to those of the L2/L3 control disc (p < 0.05, Fig. 9A).

When changes in total histological scores over time were analyzed using the Friedman test, lumbar artery ligation had a significant (p < 0.05) effect on temporal changes of scores of ischemic, but not of control (L2/L3) discs. Compared to 0-week discs, total histological scores of caudal discs were significantly higher at 4 weeks after surgery and those of cranial, bilateral and caudal discs were significantly higher at 8 weeks after surgery. However, for all ischemic discs, no significant differences were identified at 12 weeks after surgery compared to those of 0 weeks.

When histological grading scores were evaluated for each category, the scores of the AF (category A) and the border between the AF and NP (category B) were constant at grade 1 in all discs and did not change throughout the experimental period. The histological grade of other categories including NP cellularity (category B, Fig. 9B), NP matrix (category C, Fig. 9C) and NP pericellular matrix (category D, Fig. 9D) showed no significant differences in both the inter-group comparisons and temporal changes.

There was a significant negative correlation between MRI T2-values and total histological grading scores (correlation coefficient [CC] = − 0.50, p < 0.0001, Spearman’s rank-order correlation test, Fig. 10A). A significant negative correlation was also identified between MRI T2-value and every histological parameter of NP tissues (NP cellularity: CC = − 0.34, p < 0.01; NP matrix: CC = − 0.31, p < 0.01; NP pericellular matrix: CC = − 0.46, p < 0.0001) (Fig. 10B-D).

HIF-1α immunoreactivity was clearly detected in the nuclei and cytoplasm in the AF and NP cells of both control and ischemia (cranial, bilateral and caudal) discs (Fig. 11A). Statistical analysis revealed that the percentage of HIF-1α immunoreactive cells was significantly associated both with disc area (oAF, iAF, NP) and disc level (control and ischemia discs) at 4 weeks after the ligation of lumbar arteries (p < 0.05). Analysis of variance also showed a significant interaction between the disc area and disc level (p < 0.05). A post-hoc test revealed that the percentage of HIF-1α immunoreactive cells in the NP was significantly higher than that of the iAF (NP: 85.9 ± 9.6%, iAF: 63.5 ± 19.2%, oAF: 81.2 ± 19.8%, p < 0.01, between NP and iAF) (Fig. 11B). The percentage of immunoreactive cells of caudal discs in total area (oAF, iAF and NP) was significantly higher than that of control (L2/3) discs (control discs: 71.3 ± 25.4%, cranial discs: 78.7 ± 15.8%, bilateral discs: 77.2 ± 18.3%, caudal discs: 80.2 ± 15.6%, p < 0.05 between control and caudal discs) (Fig. 11B).