Activation of Hypoxia-Inducible Factor-1α Signaling Pathway Has the Protective Effect of Intervertebral Disc Degeneration

Intervertebral discs (IVDs) have poor nutrient diffusion, because the nucleus pulposus (NP) lacks direct vascular supply and likely generates adenosine triphosphate by anaerobic glycolysis. Regulation of glycolysis is mediated by hypoxia-inducible factor-1α (HIF-1α), a transcription factor that responds to local oxygen tension. Constitutively active HIF-1α (CA HIF-1α) was created by point mutation and determined the protective role of HIF-1α in IVD degeneration. Under fluoroscopy, rat caudal IVD segments were stabbed by a needle puncture, and pcDNA3- HIF-1α wild-type (WT) or pcDNA3-CA HIF-1α was transfected into NP cell lines. The constitutive activity of CA HIF-1α was analyzed using a luciferase assay after cell lysis. Next, IVD tissue samples were retrieved from five patients with degenerative lumbar spinal stenosis at the time of surgery, and NP cells were cultured. NP cells were transfected with CA HIF-1α, and relevant gene expression was measured. HIF-1α protein levels in the nucleus were significantly higher, and transcriptional activity was 10.3-fold higher in NP cells with CA HIF-1α than in those with HIF-1α WT. Gene transfer of CA HIF-1α into NP cells enhanced the expression of Glut-1, Glut-3, aggrecan, type II collagen, and Sox9. Moreover, CA HIF-1α reduced the apoptosis of NP cells induced by the Fas ligand. The HIF-1α and collagen 2 expression levels were notably increased in the NP cells of the CA HIF-1α transfected segments in histology and immunohistochemistry study. Collectively, these results suggest that activation of HIF-1α signaling pathway may play a protective role against IVD degeneration and could be used as a future therapeutic agent.


Introduction
Progressive reduction in proteoglycan content in the nucleus pulposus (NP) that deteriorates with age can lead to intervertebral disc (IVD) degeneration [1]. IVD degeneration is one of the predominant causes of lower back pain [2]. Multiple factors, including altered mechanical loading, reduced nutrient supply, hereditary factors, and changes in the extracellular microenvironment with age, have been implicated in the initiation and progression of the degenerative cascade [3][4][5][6]. However, the mechanism underlying IVD degeneration has not been clearly identified.
Cell density associated with fully developed, healthy IVD is very low, and the poor diffusion of nutrients and accumulation of metabolic waste products, such as lactate, present IVD cells with further environmental challenges due to the absence of microvasculature [7][8][9]. That is, the NP has no direct vascular supply, and adenosine triphosphate (ATP) is likely generated by anaerobic glycolysis [10]. Regulation of glycolysis is mediated by hypoxia-inducible factor-1α (HIF-1α), a transcription factor that responds to the local oxygen tension, and it has been shown to be expressed by NP cells [11].
HIF-1α promotes the transcription of key glycolytic enzymes, allowing cells to adapt to low oxygen tension and to inhibit ATP generation through oxidative phosphorylation [12]. The activity of HIF-1α within a cell is oxygen-sensitive, decreases in response to normoxia, and increases under hypoxic conditions. Under normoxic conditions, HIF-1α is hydroxylated at two specific proline residues (Pro402 and Pro564) in the oxygendependent degradation domain by the prolyl 4-hydroxylase domain-containing (PHD) enzyme, which leads to polyubiquitination and proteasomal degradation of HIF-1α by the von Hippel-Lindau tumor suppressor (VHL) protein, a component of E3 ubiquitin-protein ligase [13,14]. Under hypoxic conditions, inhibition of PHD enzyme activity spares HIF-1α from polyubiquitination and proteasomal degradation, thereby allowing HIF-1α to accumulate and to translocate to the nucleus, where it dimerizes with HIF-1β and binds to the hypoxia-responsive element (HRE) sequences of target gene promoters. In the nucleus, the transcriptional activity of the HIF-1α heterodimer is regulated by the hydroxylase factor inhibiting HIF-1 (FIH-1) [15]. The FIH-1 enzyme hydroxylates Asn803 on HIF-1α, which represses HIF-1α transactivation by preventing the transcriptional coactivator p300/CBP from binding to the HIF-1α C-transactivation domains (C-TADs) [16,17].
Recently, several researchers have reported that the PHD/HIF-1 axis is related to IVD degeneration [18,19]. However, limited information is available regarding the role of HIF-1α in IVD degeneration, despite the well-known effects of oxygen tension on HIF-1α.
To investigate the role of HIF-1α in IVD degeneration, the availability of an experimental animal model may be essential. A recently developed needle-puncture model using a stab incision (also called "anulotomy") at the IVD to induce degeneration was used for its convenience, minimal invasiveness, and cost-effectiveness [20,21]. This procedure can produce morphological and biochemical alterations similar to those observed with human IVD degeneration. With lower costs and less precision needed, the caudal discs of a rat model using needle puncture could be an effective way to demonstrate the morphological changes associated with IVD degeneration.
In this study, we hypothesized that constitutively active HIF-1α (CA HIF-1α) has a protective role against IVD degeneration. To confirm this hypothesis, we constructed CA HIF-1α, which resisted degradation even under normoxic conditions, by point mutation, and the effect of CA HIF-1α on IVD degeneration was investigated by analyzing glycolysis, apoptosis, and vascular endothelial growth factor (VEGF) and matrix gene expression. We performed magnetic resonance imaging (MRI) and histological and immunohistochemical assessments in a rat model, which demonstrated the effect of CA HIF-1α on IVD degeneration.

Hypoxia Increases Aggrecan and Collagen II mRNA Expression in Nucleus Pulposus (NP) Cells
To assess the effect of HIF-1α on disc regeneration, NP cells were cultured in either DMSO or hypoxia-mimetic agents desferrioxamine (DFO) for 48 h, and the mRNA level was quantified using qRT-PCR. NP cells cultured under hypoxia showed more Aggrecan and Collagen II mRNA expression than those cultured in normoxia ( Figure 1).

Figure 1.
Hypoxia increases mRNA expression of Aggrecan and Collagen II in nucleus pulposus (NP) cells. The mRNA levels of Aggrecan and Collagen II in NP cells treated with DMSO or Deferoxamine (150 µM) were analyzed by qRT-PCR. β-Actin was used as a normalizer. Error bars represent mean ± SD. * p < 0.05 (n = 6).

Robust Transcriptional Activity of pHIF-1α Mut
To determine whether pHIF-1α Mut A and B were degraded under normoxic conditions, nuclear and cytoplasmic fractions were obtained, and the HIF-1α protein levels of each group were analyzed by Western blotting. HIF-1α protein levels in both pHIF-1α Mut A and B were significantly higher than those of pHIF-1α WT in the nuclear fractions (p < 0.01) (Figure 2A). This suggests that HIF-1α mutated cells have increased stability and nuclear localization even under normoxic conditions, as HIF-1α is not degraded by the ubiquitin-proteasome system.  β-Actin was used as a normalizer. Error bars represent mean ± SD. * p < 0.05 (n = 6).

Robust Transcriptional Activity of pHIF-1α Mut
To determine whether pHIF-1α Mut A and B were degraded under normoxic conditions, nuclear and cytoplasmic fractions were obtained, and the HIF-1α protein levels of each group were analyzed by Western blotting. HIF-1α protein levels in both pHIF-1α Mut A and B were significantly higher than those of pHIF-1α WT in the nuclear fractions (p < 0.01) (Figure 2A). This suggests that HIF-1α mutated cells have increased stability and nuclear localization even under normoxic conditions, as HIF-1α is not degraded by the ubiquitin-proteasome system.
Next, to evaluate the transcriptional activity of pHIF-1α Mut A and B, a luciferase reporter assay was performed. In the luciferase reporter assay performed with pGL2-EPO-HRE reporter plasmids, pHIF-1α Mut A and B showed 4.7-and 10.3-fold higher transcriptional activity, respectively, compared with pHIF-1α WT with statistical significance, while no significant difference between the transcriptional activity of pHIF-1α Mut A and B and that of pHIF-1α WT was observed in the assay performed with pGL2-EPO-mHRE ( Figure 2B). Therefore, the mutants of HIF-1α showed increased transcriptional activity in a specific manner in the EPO-HRE sequence. pHIF-1α Mut B was selected for further study because pHIF-1α Mut B was more active than pHIF-1α Mut A. A pHIF-1α Mut B will be referred to as CA HIF-1α from this point.
Then, to investigate the effects of CA HIF-1α (pHIF-1α Mut B) on VEGF and glucose transporter (Glut-1 and Glut-3) expression, NP cells from patient tissues were transfected with pcDNA3, pHIF-1α WT, and CA HIF-1α. After 48 h, qRT-PCR analysis was performed to evaluate VEGF, Glut-1, and Glut-3 expression. CA HIF-1α induced significant upregulation of VEGF, Glut-1, and Glut-3 expression in NP cells. To validate the effect of CA HIF-1α on the disc matrix, Sox9, aggrecan, and type II collagen expression were also analyzed, and the results indicated that CA HIF-1α increased Sox9, aggrecan, and type II collagen expression in NP cells ( Figure 2C).
To determine whether pHIF-1α Mut A and B were degraded under normoxic condi-tions, nuclear and cytoplasmic fractions were obtained, and the HIF-1α protein levels of each group were analyzed by Western blotting. HIF-1α protein levels in both pHIF-1α Mut A and B were significantly higher than those of pHIF-1α WT in the nuclear fractions (p < 0.01) (Figure 2A). This suggests that HIF-1α mutated cells have increased stability and nuclear localization even under normoxic conditions, as HIF-1α is not degraded by the ubiquitin-proteasome system.  were transfected into NP cells. After 2 days, nuclear and cytoplasmic fractions were obtained, and the HIF-1α protein levels were examined by Western blotting. A pHIF-1α Mut A and B had significantly higher HIF-1α protein levels than pHIF-1α WT in both nuclear and cytoplasmic fractions. DFO was used as a positive control. Lamin A/C was used as nuclear control. α-Tubulin was used as nuclear control. (B) CA HIF-1α (pHIF-1α Mut B) enhances HIF-1α transcriptional activity. A pHIF-1α Mut A and B have higher transcriptional activity specifically related to the HRE sequence. DFO only, pcDNA3 only, pHIF-1α WT, pHIF-1α Mut A, and pHIF-1α Mut B were transfected into NP cells with pGL2-EPO-HRE or pGL2-EPO-mHRE under normoxic condition. After 2 days, cells were lysed, and luciferase activities measured. Both mutants of HIF-1α showed increased transcriptional activity specifically related to the EPO-HRE sequence. A pHIF-1α Mut B was more active than Mut A. Firefly luciferase activity was normalized to that of Renilla luciferase, which was used as an internal control. * and ** mean p-value < 0.05 and <0.01 comparison within the pGL2-EPO-HRE groups, respectively. # and ## mean p-value < 0.05 and <0.01, respectively, between pGL2-EPO-HRE and pGL2-EPO-mHRE in the same group. (C) CA HIF-1α (pHIF-1α Mut B) upregulates the genes related to the metabolic activity and extracellular matrix production. NP cells from patient tissues were treated with control pcDNA3 or CA HIF-1α. After 48 h, qRT-PCR analysis was performed. (Left) CA HIF-1α induced significant upregulation of VEGF, Glut-1, and Glut-3 expression in NP cells. (Right) CA HIF-1α increased Sox9, aggrecan, and type II collagen expression in NP cells with the statistical significance. The total mRNA quantity was normalized to that of GAPDH. * p-value < 0.05.

CA HIF-1α Decreased the Apoptosis of NP Cells
To examine the role of CA HIF-1α in NP cell apoptosis, NP cells were transfected with pHIF-1α WT or CA HIF-1α, and apoptosis was induced by treatment with 50 ng/mL FasL. After 24 h, apoptosis was measured by enumerating annexin V-positive cells. FasL induced apoptosis in 34.7% of NP cells, although CA HIF-1α reduced the proportion of apoptotic cells to 12.5% ( Figure 3). This result implies that CA HIF-1α protects NP cells from apoptosis during IVD degeneration.

Radiographic and MRI Assessment, and Histological Evaluation of CA HIF-1α
The T2 signal intensity was decreased in the PBS-and pcDNA3-alone segments at 4 and 8 weeks following needle puncture. However, the signal intensity in the pHIF-1α WT and pHIF-1α Mut B segments was increased at 4 and 8 weeks following injection compared with that of the PBS and pcDNA3 alone segments, and the difference was statistically significant (p < 0.05; Figure 4A). The degree of disc height index (DHI) reduction comparing control in the pHIF-1α WT and pHIF-1α Mut B segments was smaller than that in the PBS and pcDNA3 segments at 4 and 8 weeks after injection; this difference was also statistically significant (p < 0.05; Figure 4B).
Representative images of safranin O staining are shown in Figure 4C. The NP area in the disc stained strongly for safranin O, although the stained levels were not homogenous in the PBS and pcDNA3 segments. Histological analysis of IVD tissues was performed using H&E and toluidine blue staining. In the PBS and pcDNA3 segments, ruptured and serpentine-patterned fibers were observed in the AF. Furthermore, the border between the AF and NP was interrupted and unclear and exhibited features of severe degeneration. However, the AF and the border between the AF and NP in the pHIF-1α WT and pHIF-1α Mut B segments demonstrated an intact circumferential AF, and the border between the AF and NF was clear compared with those observed in the PBS and pcDNA3 segments. Immunohistochemical staining revealed that the staining for HIF-1α and collagen 2 were stronger in the pHIF-1α Mut B transfected segments compared with pHIF-1α WT segments. HIF-1α and collagen 2 expression levels were negligible in the PBS segments. Conversely, the HIF-1α and collagen 2 expression levels were notably increased in the NP cells of the pHIF-1α Mut B transfected segments. A histopathology scoring system was applied according to Lai et al. [22], and pHIF-1α Mut B had a significantly lower score than the other segments.

Radiographic and MRI Assessment, and Histological Evaluation of CA HIF-1α
The T2 signal intensity was decreased in the PBS-and pcDNA3-alone segments at 4 and 8 weeks following needle puncture. However, the signal intensity in the pHIF-1α WT and pHIF-1α Mut B segments was increased at 4 and 8 weeks following injection compared with that of the PBS and pcDNA3 alone segments, and the difference was statistically significant (p < 0.05; Figure 4A). The degree of disc height index (DHI) reduction comparing control in the pHIF-1α WT and pHIF-1α Mut B segments was smaller than that in the PBS and pcDNA3 segments at 4 and 8 weeks after injection; this difference was also statistically significant (p < 0.05; Figure 4B).
Representative images of safranin O staining are shown in Figure 4C. The NP area in the disc stained strongly for safranin O, although the stained levels were not homogenous in the PBS and pcDNA3 segments. Histological analysis of IVD tissues was performed using H&E and toluidine blue staining. In the PBS and pcDNA3 segments, ruptured and serpentine-patterned fibers were observed in the AF. Furthermore, the border between the AF and NP was interrupted and unclear and exhibited features of severe degeneration. However, the AF and the border between the AF and NP in the pHIF-1α WT and pHIF-1α Mut B segments demonstrated an intact circumferential AF, and the border between the AF and NF was clear compared with those observed in the PBS and pcDNA3 segments. Immunohistochemical staining revealed that the staining for HIF-1α and collagen 2 were stronger in the pHIF-1α Mut B transfected segments compared with pHIF-1α WT segments. HIF-1α and collagen 2 expression levels were negligible in the PBS segments. Conversely, the HIF-1α and collagen 2 expression levels were notably increased in the NP cells of the pHIF-1α Mut B transfected segments. A histopathology scoring system was applied according to Lai et al. [22], and pHIF-1α Mut B had a significantly lower score than the other segments.

Discussion
In this study, mutants of HIF-1α were constructed by replacing Pro402, Pro564, and Asn803 with alanines to investigate the protective role of HIF-1α in IVD degeneration. HIF-1α was first identified as a critical factor for the inducible activity of the EPO 3′ enhancer and is now recognized to be a key regulator of gene expression in response to changes in cellular oxygen tension [23]. The N-terminal half of HIF-1α contains basic helix-loop-helix and Per-aryl hydrocarbon receptor nuclear translocator-Sim, which are required for dimerization and DNA binding [24]. The C-terminal half contains domains required for degradation and transactivation, the oxygen-dependent degradation domain, which confers oxygen-dependent instability, two independent TADs, and an inhibitory domain that negatively regulates TAD [25]. In addition, we demonstrated that pHIF-1α Mut B injection significantly alleviated IVD degeneration using a needle puncture model in rats, as determined by MRI and histological and immunohistochemical analyses.
Our results demonstrated that CA HIF-1α had a significantly increased protein level and nuclear localization when transfected into NP cells. This result indicates that CA HIF-1α is not degraded under normoxic conditions and indeed is constitutively active. Moreover, CA HIF-1α transactivates reporter gene expression specifically in the EPO-HRE sequence. Therefore, these results demonstrate that CA HIF-1α can also be constructed by point mutation, not by deletion as reported previously, and may be a useful tool for further investigating the biological roles of HIF-1α under normoxic conditions [26,27].
In NP cells transfected with CA HIF-1α, the expression of Glut-1 and Glut-3, glucose transporters in NP cells, was significantly enhanced under normoxic conditions. This suggests that CA HIF-1α facilitates energy metabolism in NP cells by stimulating anaerobic

Discussion
In this study, mutants of HIF-1α were constructed by replacing Pro402, Pro564, and Asn803 with alanines to investigate the protective role of HIF-1α in IVD degeneration. HIF-1α was first identified as a critical factor for the inducible activity of the EPO 3 enhancer and is now recognized to be a key regulator of gene expression in response to changes in cellular oxygen tension [23]. The N-terminal half of HIF-1α contains basic helix-loop-helix and Per-aryl hydrocarbon receptor nuclear translocator-Sim, which are required for dimerization and DNA binding [24]. The C-terminal half contains domains required for degradation and transactivation, the oxygen-dependent degradation domain, which confers oxygen-dependent instability, two independent TADs, and an inhibitory domain that negatively regulates TAD [25]. In addition, we demonstrated that pHIF-1α Mut B injection significantly alleviated IVD degeneration using a needle puncture model in rats, as determined by MRI and histological and immunohistochemical analyses.
Our results demonstrated that CA HIF-1α had a significantly increased protein level and nuclear localization when transfected into NP cells. This result indicates that CA HIF-1α is not degraded under normoxic conditions and indeed is constitutively active. Moreover, CA HIF-1α transactivates reporter gene expression specifically in the EPO-HRE sequence. Therefore, these results demonstrate that CA HIF-1α can also be constructed by point mutation, not by deletion as reported previously, and may be a useful tool for further investigating the biological roles of HIF-1α under normoxic conditions [26,27].
In NP cells transfected with CA HIF-1α, the expression of Glut-1 and Glut-3, glucose transporters in NP cells, was significantly enhanced under normoxic conditions. This suggests that CA HIF-1α facilitates energy metabolism in NP cells by stimulating anaerobic glycolysis. Degeneration is characterized by increased degradation of the normal IVD matrix by locally produced matrix metalloproteinases and a disintegrin and metalloproteinase with thrombospondin motifs [28]. In addition, the matrix produced in degenerated IVDs differs from that in normal IVDs, as a consequence of switches in the production of collagen within the inner annulus fibrosus and NP from type II to type I, and in the synthesis of proteoglycan from aggrecan to versican, biglycan, and decorin [29,30]. The resultant changes within the extracellular matrix have several consequences, including loss of structural integrity, decreased hydration, and reduced ability to withstand load.
Histologically, in young adults, NP contains a high percentage of type II collagen synthesized by disc cells. With aging, the synthesis of type II collagen declines, whereas the degradation of existing type II collagen gradually increases [31,32]. Furthermore, type I collagen production within the disc increases, leading to a higher proportion of less compliant type I collagen [33].
The decline in type II collagen within NP features prominently in the pathogenesis of IVD degeneration, and reversing this trend may provide an opportunity to modify early IVD degeneration. The regulation of type II collagen synthesis undoubtedly involves a complex interplay between local and systemic factors. Recent studies have demonstrated that Sox9 is an essential transcription factor for type II collagen synthesis and chondrogenesis [34][35][36]. Aggrecan and type II collagen, major components of proteoglycans and the extracellular matrix of NP cells, play a role in spinal stability. Sox9 is a representative chondrogenesis-inducing transcription factor. In this study, Sox9, aggrecan, and type II collagen expression increased in NP cells transfected with CA HIF-1α. This suggests that transfer of the CA HIF-1α gene into NP cells may stimulate extracellular matrix synthesis and play a protective role in IVD degeneration, because enhanced matrix degradation is a hallmark of IVD degeneration.
In degenerative IVD tissues, increased cell apoptosis and high expression of both Fas and FasL have been reported [37,38]. In this study, apoptosis of NP cells was induced by treatment with FasL, and apoptosis was markedly reduced in NP cells transfected with HIF-1α. Therefore, CA HIF-1α gene transfer into NP cells may decrease apoptosis and protect against IVD degeneration.
We have, for the first time, identified an association between IDD and the CA HIF-1α, especially pHIF-1α Mut B. Sutter et al. demonstrated that missense mutations increase HIF-1α expression under normoxic conditions by blocking ubiquitination and missense mutations and/or deletions involving several different regions of HIF-1α, resulting in constitutive activation and transcriptional activity in normoxic cells.
We utilized MRI technology to assess IVD degeneration intravitally and continuously in vivo. The signal intensity of the T2-weighted images and disc height on the MR scans showed that the water content and NP volume in the IVD as well as the loss of water content and NP volume were associated with a decrease in the T2-weighted signal intensity and disc height. Therefore, IVD regeneration results in an increase in T2-weighted signal intensity and disc height. Our results demonstrated that degenerative changes in NP induced by needle puncture were recovered following pHIF-1α Mut injection, especially in pHIF-1α Mut B. The T2-weighted signal intensity and disc height of the pHIF-1α Mut B were significantly higher than those of the other groups at 4 and 8 weeks after injection.
The degenerated disc was less elastic, consequently prohibiting the ability to absorb and to endure spinal loading [16]. Type II collagen in IVD was demonstrated to be present at normal levels in the NP and in the inner layer of the AF; however, type I collagen was found to be present at normal levels in the AF and degenerative NP [39]. In a rabbit annular stab-injury model, aggrecan and type IIa collagen mRNA levels were decreased and could not be restored to normal levels, whereas type Ia collagen mRNA levels gradually increased throughout the course of degeneration [20]. We recognized that this rat model has limitation, because we used the multiple discs within the same animal. Despite this, this study demonstrated that in degenerative discs injected with pHIF-1α Mut B, VEGF, aggrecan, and collagen II levels in the NP were restored. Our results indicate that CA HIF-1α may be an optimal method for maintaining the normal extracellular matrix components and mechanical properties of discs.
In summary, we successfully constructed HIF-1α mutant form, which was not degraded under normoxic conditions by point mutation. CA HIF-1α showed approximately 10.3-fold higher transcriptional activity than HIF-1α WT from the EPO-HRE reporter plasmid. Gene transfer of CA HIF-1α into NP cells using a recombinant adenovirus enhances the expression of glycolysis-involved genes and matrix synthesis-related genes. Furthermore, CA HIF-1α reduced apoptosis of NP cells. Using MRI, histological, and immunohistochemical analyses, we found that CA HIF-1α inhibited IVD degeneration and maintained NP in the disc of rat model.
Collectively, these results suggest that activation of the HIF-1α signaling pathway may play a protective role against IVD degeneration and could be used as a future therapeutic agent.

Validation of HIF-1α Mut by Western Blot Analysis
To compare the activity of HIF-1α WT proteins between mutants A and B and the control under normoxic conditions (20% O 2 ), the following five groups were transfected into NP cells using Lipofectamine (Invitrogen, Waltham, MA, USA): (1) Desferrioxamine (DFO) only, (2) pcDNA3 only, (3) pHIF-1α WT, (4) pHIF-1α Mut A, and (5) pHIF-1α Mut B. DFO was used as a positive control, because it inhibits both PHD enzyme and FIH-1 enzyme activities [41]. The transfected cells were lysed using a cell fractionation kit (Cell Signaling Technology, Danvers, MA, USA). The cells were separated into nuclear and cytoplasmic fractions in Nupage 4-12% Bis-Tris gels (Invitrogen, Waltham, MA, USA) and then transferred to polyvinylidene difluoride membranes (Millipore, Billerica, MA, USA). The membranes were blocked with 5% skim milk and washed with phosphate-buffered saline and 0.2% Tween 20 buffer. The membranes were blotted with mouse anti-HIF-1α (Cell Signaling Technology, Danvers, MA, USA), mouse anti-nuclear lamin A/C, or anti-αtubulin (Santa Cruz Biotechnology, Dallas, TX, USA) antibodies. The blots were treated with goat anti-mouse conjugated with horseradish peroxidase (Santa Cruz Biotechnology, Dallas, TX, USA) and visualized using an enhanced chemiluminescence system (Uppsala, Sweden).

Validation of pHIF-1α Mut by Luciferase Reporter Assay
To determine the transcriptional activity of pHIF-1α Mut A and B, a luciferase reporter assay was performed. The following two different types of luciferase reporter plasmids were used to determine whether pHIF-1α Mut A and B had increased transcriptional activity specifically related to the erythropoietin (EPO)-HRE sequence: (1) pGL2-EPO-HRE, containing the wild-type HRE sequence in the EPO gene, and (2) pGL2-EPO-mHRE, containing the mutant HRE sequence in the EPO gene. DFO only, pcDNA3 only, pHIF-1α WT, and pHIF-1α Mut A and B were co-transfected into HEK NP cells with pGL2-EPO-HRE or pGL2-EPO-mHRE using Lipofectamine (Invitrogen, Waltham, MA, USA). The cells were lysed, and luciferase assays (Promega, Madison, WI, USA) were performed using a luminometer (Titertek Berthold FB14, Bad Wildbad, Germany). Firefly luciferase activity was normalized to that of Renilla luciferase, which was used as an internal control.

Patients and Tissue Samples
The Institutional Review Board of the CHA Bundang Medical Center approved this study. Patients diagnosed with degenerative lumbar spinal stenosis were recruited after obtaining informed consent. Before sample collection, the degree of IVD degeneration was evaluated according to the Pfirrman grading system [42], and degenerative discs of grade IV or higher were collected in this study.

Human NP Cell Isolation and Cultures
Human NP cells donated by patients were obtained within 2 h after surgery. NP tissues were identified and separated by a stereotaxic microscope. The NP tissues were washed with phosphate-buffered saline (PBS) and digested for 40 min in 0.2% pronase (Gibco BRL, Carlsbad, CA, USA). Following being washed, the tissues were incubated in 0.25% type II collagenase (Gibco BRL, Carlsbad, CA, USA) at 37 • C under gentle agitation for 4 h. The cell suspension was centrifuged at 300× g for 10 min, and the pellet was washed with Dulbecco's modified Eagle's medium (DMEM) (Invitrogen, Waltham, MA, USA). The cells were plated in culture dishes and maintained in DMEM supplemented with 10% fetal bovine serum and 50 µg/mL gentamicin (Invitrogen, Waltham, MA, USA) at 37 • C in 5% CO 2 . First-passage cells maintained in a monolayer were obtained, and second-passage cells were used in the experiments. NP cells were transfected with empty pcDNA3 only or pHIF-1α (WT, Mut A and Mut B).

Quantitative Real-Time Polymerase Chain Reaction (qRT-PCR)
Total RNA was isolated from NP cells using TRIzol reagent (Invitrogen, Waltham, MA, USA). First-strand cDNA was synthesized using the SuperScript system (Invitrogen, Waltham, MA, USA). PCR amplification conditions were as follows: denaturation at 94 • C for 20 s, annealing at 55 • C for 30 s, and extension at 72 • C for all genes. PCR reactions were performed using the QuantiTech SYBR Green PCR kit (Qiagen, Valencia, CA, USA) and an Exicycler 96 system (Bioneer, Daejeon, Korea). Specific primers for the genes examined were based on their PrimerBank sequences and are listed in Table 1.

Preparation of Subcellular Fractions
NP cells were cultured and fractionated into cytosolic and nuclear fractions using a cell fractionation kit (Abcam, Inc., Cambridge, UK) according to the manufacturer's instructions. Trypsinized NP cells were collected and centrifuged for 5 min at 300× g. The collected cells were washed with buffer A (wash buffer). Then, the cells were resuspended in buffer A. An equal volume of buffer B (lysis buffer for Cyt extraction) was added to the cell suspensions, mixed via pipetting, and then incubated for 7 min at room temperature (RT). The cell lysates were centrifuged at 5000× g for 1 min at 4 • C, and the cytosol fractions (cell lysate supernatants) were collected in new tubes. The remaining pellets containing the nuclear protein fraction were resuspended in buffer A.

Western Blotting
Transfected NP cell lysates were boiled in an SDS sample buffer, resolved by SDS-PAGE, and transferred to nitrocellulose membranes. After the transfer, the membranes were blocked in 5% skim milk in TBST (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.05% Tween 20) for 40 min and incubated with the specific primary antibodies in the blocking solution for overnight at 4 • C. Antibodies against HIF-1α (sc-13515, 1:1000) and α-Tubulin (sc-8035, 1:1000) were purchased from Santa Cruz Biotechnology (Dallas, TX, USA). Antibodies against Lamin A/C (#4777, 1:1000) was purchased from Cell Signaling Technology (Danvers, MA, USA). The membranes were then washed with TBST and incubated with horseradish peroxidase (HRP)-conjugated secondary antibody followed by detection using an enhanced chemiluminescence system (Amersham Pharmacia Biotech, Piscataway, NJ, USA).

Preparation of Animals
Ten male Sprague Dawley rats (each weighing 220-250 g and 2 months old) were used in this study. The animals were housed with free access to commercial rodent chow and water. The temperature was maintained at 24 • C, and the light schedule comprised 12 h of daylight starting at 8:00 a.m. and 12 h of darkness starting at 8:00 p.m. Humidity was maintained at 50%. All experiments were performed according to the guidelines for the ethical treatment of animals approved by the Laboratory Animal Ethics Committee of CHA Bundang Medical Center (Seongnam, Korea).

Surgical Procedure
The rat tail segments were divided into four groups: (1) phosphate-buffered saline (PBS), (2) pcDNA3 only, (3) pHIF-1α WT, and (4) pHIF-1α Mut B. Ten rat tails were prepared for the drug treatment. The animals were anesthetized by inhalation of anesthetic isoflurane, and the operative field was prepared in a sterile fashion. After identifying the level of the caudal segments by fluoroscopy, a puncture was made parallel to the endplates through the annulus fibrosus (AF) into the NP using a syringe needle (21-gauge) ( Figure 5) [43]. The needle was then rotated by 180 • and held for 5 s. During needle puncture, fluoroscopy was performed to ensure that the needle reached the center of the NP [20]. For the MRI assessment, the caudal segments 4/5 and 5/6 from the same rat tail were used for the PBS-and pcDNA3-only treatment, and the caudal segments 6/7 and 7/8 for the pHIF-1α WT and pHIF-1α Mut B ( Figure 5).

Radiographic and MRI Evaluation
Images were obtained using a 3.0T MRI machine (Philips Healthcare, Amster The Netherlands) using a dedicated coil for small animals. The animals were seriall lowed by MRI immediately and 4 and 8 weeks following injection. At each time poin rats were anesthetized with a 2% isoflurane/oxygen mixture throughout the MRI ex nation. The animals were placed in a prone position, and their tails were straighten the MR scanner. Serial T2-weighted sagittal and transverse images covering the entir perimental disc area were obtained, and T2-weighted signal intensity and disc h were measured using Image J software (National Institutes of Health, Bethesda, USA). Additionally, intervertebral disc height index (DHI) under X-ray immediatel fore and after injection was calculated using modification of previous methods [44 Image assessments were conducted by two independent, blinded, and experienced servers. The data are presented as the means of the two evaluations.

Histological Analysis
At 8 weeks after injection, all rats were killed, and the discs were harvested fo tological examination. Each disc was fixed, decalcified, using 10% ethylenediamin raacetic acid (EDTA; pH = 7.4) for 1 month. The EDTA solution was replaced every 2 The decalcified discs were washed, dehydrated and then processed for paraffin em ding and sectioning into sagittal sections (5 µm thickness) using a microtome. Sa sections were stained with hematoxylin and eosin (H&E), safranin O, and toluidine Histological images were analyzed qualitatively under a light microscope (BX51, O pus, Inc., Tokyo, Japan) at magnifications ranging from 4 to 100× to investigate chang the NP, AF, and endplates. The extent of disc degeneration was graded using a h index as described previously by Masuda et al. [47].

Immunohistochemical Staining
Immunohistochemistry (IHC) was performed using a Ready-to-use IHC/ICC kit Vision, Inc., Milpitas, CA, USA) according to the manufacturer's protocol. Briefly, p fin-embedded sections were deparaffinized, rehydrated, immersed in a retrieval sol (10 mmol/L citrate, pH 6.0), and then placed in a microwave for 10 min. Endogenou roxidase activity was blocked using 3% hydrogen peroxide for 10 min. The slides incubated in 3% H2O2 at room temperature for 30 min to quench endogenous perox activity and then blocked in blocking buffer (BioVision, Inc., Milpitas, CA, USA) at temperature for 10 min, followed by incubation with anti-HIF-1α (sc-13515; 1:100) or collagen II (1:100) antibodies at room temperature for 30 min. After incubation with H anti-mouse or -rabbit IgG polymer at room temperature for 20 min and washing with

Radiographic and MRI Evaluation
Images were obtained using a 3.0T MRI machine (Philips Healthcare, Amsterdam, The Netherlands) using a dedicated coil for small animals. The animals were serially followed by MRI immediately and 4 and 8 weeks following injection. At each time point, the rats were anesthetized with a 2% isoflurane/oxygen mixture throughout the MRI examination. The animals were placed in a prone position, and their tails were straightened in the MR scanner. Serial T2-weighted sagittal and transverse images covering the entire experimental disc area were obtained, and T2-weighted signal intensity and disc height were measured using Image J software (National Institutes of Health, Bethesda, MD, USA). Additionally, intervertebral disc height index (DHI) under X-ray immediately before and after injection was calculated using modification of previous methods [44][45][46]. Image assessments were conducted by two independent, blinded, and experienced observers. The data are presented as the means of the two evaluations.

Histological Analysis
At 8 weeks after injection, all rats were killed, and the discs were harvested for histological examination. Each disc was fixed, decalcified, using 10% ethylenediaminetetraacetic acid (EDTA; pH = 7.4) for 1 month. The EDTA solution was replaced every 2 days. The decalcified discs were washed, dehydrated and then processed for paraffin embedding and sectioning into sagittal sections (5 µm thickness) using a microtome. Sagittal sections were stained with hematoxylin and eosin (H&E), safranin O, and toluidine blue. Histological images were analyzed qualitatively under a light microscope (BX51, Olympus, Inc., Tokyo, Japan) at magnifications ranging from 4 to 100× to investigate changes in the NP, AF, and endplates. The extent of disc degeneration was graded using a height index as described previously by Masuda et al. [47].

Immunohistochemical Staining
Immunohistochemistry (IHC) was performed using a Ready-to-use IHC/ICC kit (BioVision, Inc., Milpitas, CA, USA) according to the manufacturer's protocol. Briefly, paraffin-embedded sections were deparaffinized, rehydrated, immersed in a retrieval solution (10 mmol/L citrate, pH 6.0), and then placed in a microwave for 10 min. Endogenous peroxidase activity was blocked using 3% hydrogen peroxide for 10 min. The slides were incubated in 3% H2O2 at room temperature for 30 min to quench endogenous peroxidase activity and then blocked in blocking buffer (BioVision, Inc., Milpitas, CA, USA) at room temperature for 10 min, followed by incubation with anti-HIF-1α (sc-13515; 1:100) or anti-collagen II (1:100) antibodies at room temperature for 30 min. After incubation with HRP-anti-mouse or -rabbit IgG polymer at room temperature for 20 min and washing with PBS, the tissue sections were treated with 3,3 -diaminobenzidine at room temperature for 10 min, followed by counterstaining with hematoxylin at room temperature for 1 min.