Normoxic stabilization of HIF-1α drives glycolytic metabolism and regulates aggrecan gene expression in nucleus pulposus cells of the rat intervertebral disk

Amit Agrawal, Asha Guttapalli, Srinivas Narayan, Todd J. Albert, Irving M. Shapiro, Makarand V. Risbud


The nucleus pulposus is an aggrecan-rich, avascular tissue that permits the intervertebral disk to resist compressive loads. In the disk, nucleus pulposus cells express hypoxia-inducible factor (HIF)-1α, a transcription factor that responds to oxygen tension and regulates glycolysis. The goal of the present study was to examine the importance of HIF-1α in rat nucleus pulposus cells and to probe the function of this transcription factor in terms of regulating aggrecan gene expression. We found that HIF-1α protein levels and mRNA stability were similar at 20 and 2% O2; there was a small, but significant increase in HIF-1α transactivation domain activity in hypoxia. With respect to HIF-1α target genes GAPDH, GLUT-1, and GLUT-3, mRNA and protein levels were independent of the oxygen tension. Other than a modest increase in 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase reporter activity, the oxemic state did not change GAPDH, GLUT-1, and GLUT-3 promoter activities. Treatment of cells with 2-deoxyglucose (2-DG), a glycolytic inhibitor, resulted in a significant suppression in ATP synthesis in normoxia, whereas treatment with mitochondrial inhibitors did not affect ATP production and cell viability. However, measurement of the rate of fatty acid oxidation indicated that these cells contained functioning mitochondria. Finally, we showed that when HIF-1α was suppressed, irrespective of the oxemic state, there was a partial loss of aggrecan expression and promoter activity. Moreover, when cells were treated with 2-DG, there was inhibition in aggrecan promoter activity. Results of this study indicate that oxygen-independent stabilization of HIF-1α in nucleus pulposus cells is a metabolic adaptation that drives glycolysis and aggrecan expression.

  • hypoxia
  • hypoxia-inducible factor-1α

the local oxygen tension promotes cellular differentiation and regulates tissue energy metabolism. When the oxygen tension is low, there is almost complete reliance on glycolysis to generate ATP and reducing equivalents (12, 29, 30). In specialized tissues such as the intervertebral disk, the mechanism by which cells conserve energy has received little attention. Nonetheless, it is known that in the intervertebral disk, the nucleus pulposus has no direct vascular supply (8, 24, 33), and ATP is probably generated by anaerobic glycolysis (1, 10). The dependence on glycolysis reflects the low oxygen tension in the disk, which has been reported to be low as low as 2%, and the observation that modulation in the oxygen supply can cause vertebral defects (15). Regulation of glycolysis is mediated by hypoxia-inducible factor (HIF)-1α, a transcription factor that responds to the local oxygen tension, and has been shown to be expressed by nucleus pulposus cells in vivo (20).

HIF-1α is a member of the basic helix-loop-helix (bHLH)-PER-ARNT-SIM (PAS) family of proteins. It consists of a constitutively expressed β-subunit and an α-subunit that is rapidly degraded under normoxic conditions (34). The bHLH and PAS motifs are required for dimerization, and two transactivation domains (TAD), N-TAD (aa 531-575) and C-TAD (aa 786-826) bind several coactivators; whereas the downstream basic region affords specific binding to the hypoxia-response element (HRE) in target genes (9, 27). HIF-1α promotes transcription of key glycolytic enzymes, allowing a cell to adapt to low Po2 (6, 16, 21, 26, 28), and inhibits ATP generation through oxidative phosphorylation (17). In an earlier investigation, our group (22) showed that there is evidence of normoxic stabilization of this otherwise oxygen labile protein in nucleus pulposus cells.

The goal of the present study was to examine the importance of normoxic stabilization of HIF-1α in rat nucleus pulposus cells and to probe the function of this transcription factor in terms of aggrecan expression. We focused on aggrecan, since this major extracellular matrix proteoglycan permits the intervertebral disk to withstand compressive loads. We have shown that unlike other mammalian cells, under normoxic conditions, HIF-1α maintains ATP generation through glycolysis and regulates aggrecan gene expression. This finding lends strength to the view that cells of the nucleus pulposus are functionally adapted to their avascular microenvironmental niche.


Isolation of nucleus pulposus cells.

Male Wistar rats (225–250 g) were euthanized with CO2. The spinal columns were removed under aseptic conditions, and lumbar intervertebral disks were separated. The gel-like nucleus pulposus was separated from the annulus fibrosus with the use of a dissecting microscope and was treated with 0.1% collagenase and 10 U/ml hyaluronidase for 4–6 h. The partially digested nucleus pulposus tissue was maintained as an explant in Dulbecco's modified Eagle's medium (DMEM) and 10% fetal bovine serum (FBS) supplemented with antibiotics in a humidified atmosphere containing 5% CO2 at 37°C. Nucleus pulposus cells migrated out of the explant after 1 wk. When confluent, the cells were lifted using trypsin (0.25%)-EDTA (1 mM) solution and subcultured in 10-cm dishes. For some experiments, human disk tissue (minimally degenerative and regarded as surgical waste) was obtained from patients undergoing cervical spinal surgeries by following a protocol approved by the Institutional Review Board of the Thomas Jefferson University.

Cell culture in hypoxia.

Nucleus pulposus cells were cultured in an Hypoxia Work Station (Invivo2 300; Ruskinn, Brigend, UK) with a mixture of 2% O2, 5% CO2, and 93% N2 for 24–72 h. The concentration of oxygen chosen for this study was selected on the basis of our previous in vitro studies and information on the oxemic status of the disk in vivo. To induce chemical hypoxia, cells were treated with 130 μM desferrioxamine (Dfx) under normoxic conditions for 24 h. To chemically suppress HIF-1α, cells were treated with 500 mM 17-allylamino-17-demethoxygeldanamycin (17-AAG), a known HIF-1α inhibitor. In some experiments, HeLa cells were used as control and maintained in culture as described above.

Examination of HIF-1α and GLUT-1 expression by immunofluorescence analysis.

Cells were plated in flat-bottom 96-well plates and maintained in normoxia or hypoxia for 24 h. After incubation, cells were fixed with 4% paraformaldehyde on ice for 20 min and washed with phosphate-buffered saline (PBS). Cells were permeabilized with 0.5% Triton X-100 in PBS for 15 min and incubated in blocking solution (PBS containing 10% FBS) for 45 min. Cells were then incubated with anti-HIF-1α antibody (ABR, Golden, CO) or anti-GLUT-1 antibody (Abcam, Boston, MA) in blocking solution at a dilution of 1:200 at 4°C overnight. As a negative control, cells were treated with mouse isotype IgG under similar conditions. After cells were thoroughly washed with PBS, the experimental and control cells were incubated with Alexa Fluor488-conjugated anti-mouse secondary antibody (Molecular Probes, Eugene, OR) at a dilution of 1:50 for 1 h at room temperature. After washing, some cells were also incubated with 5 μM propidium iodide for 20 min at room temperature. Cells were imaged using a laser scanning confocal microscope (Olympus Fluoview, Tokyo, Japan).

Nuclear and membrane protein extraction.

Nuclear extracts were prepared from subconfluent cells according to the method of Dignam et al. (5) using the CellLytic NuCLEAR extraction kit (Sigma Aldrich, St. Louis, MO). Following culture in hypoxia or normoxia, cells were immediately placed on ice and lysed in hypotonic buffer (10 mM HEPES, pH 7.9, 1.5 mM MgCl2, 10 mM KCl, and 0.5 mM DTT) and incubated on ice for 15 min. Igepal CA-630 was added to a final concentration of 0.6%, and the mixture was vortexed vigorously for 10 s. Nuclei were recovered by centrifugation at 3,300 g for 30 s at 4°C and extracted by gentle shaking in buffer containing 20 mM HEPES, pH 7.9, 0.42 M NaCl, 25% glycerol, 1.5 mM MgC12, 0.2 mM EDTA, and 0.5 mM DTT for 30 min at 4°C. The extract was then centrifuged for 15 min at 25,000 g, and the supernatant was snap frozen at −70°C. All buffers contained a protease inhibitor cocktail [2 mM 4-(2-aminoethyl)benzenesulfonylfluoride, 1.4 pM trans-epoxysuccinyl-l-leucylamido(4-guanidino)butane, 130 pM bestatin, 1 μM leupeptin, and 0.3 pM aprotinin] and phosphatase inhibitors (5 mM NaF and 200 μM Na3VO4). Membrane proteins were isolated using the mammalian membrane protein extraction reagent kit Mem-Per (Pierce, Rockford, IL) following the manufacturer's protocol.

Western blot analysis.

Proteins were resolved on 10% SDS-polyacrylamide gels and transferred by electroblotting to nitrocellulose membranes (Bio-Rad, Hercules, CA). The membranes were blocked with 5% nonfat dry milk in TBST (50 mM Tris, pH 7.6, 150 mM NaCl, 0.1% Tween 20) and incubated overnight at 4°C in 3% nonfat dry milk in TBST with the antibodies against HIF-1α (1:500; R&D Systems, Minneapolis, MN), GLUT-1 (1:500), GLUT-3 (1:500; Abcam), or GAPDH (1:3,000; Novus, Littleton, CO) and β-actin (1:4,000) (Sigma). Immunolabeling was detected using the ECL reagent (Amersham Biosciences). Relative expression levels were determined by quantitative densitometric analysis using Kodak 1D image analysis software (Kodak, Rochester, NY).

Real-time RT-PCR analysis.

At the end of each treatment period, total RNA was extracted from cells using Trizol (Invitrogen) according to the manufacturer's protocol. RNA was further treated with RNase-free DNase I and passed through RNeasy mini columns (Qiagen). Total RNA (100 ng) was used as template for real-time PCR analysis. Reactions were set up in microcapillary tubes using 1 μl of RNA with 9 μl of a LightCycler FastStart DNA Master SYBR green I mix (Roche Diagnostics, Indianapolis, IN) to which gene-specific forward and reverse PCR primers were added. With each set of samples, no template control was included. PCR reactions were performed in LightCycler (Roche) according to the manufacturer's instructions. All the primers used were synthesized by Integrated DNA Technologies (Coralville, IA). See Table I for sequences.

View this table:
Table 1.

Primer sequences used for real-time RT-PCR analysis

Reporter plasmids.

Luciferase reporter plasmids were provided by Dr. Sherin Devaskar, University of California, Los Angels (Glut-3 promoter) (19), Dr. Oleksandr Minchenko, Palladin Institute of Biochemistry, Ukraine [6-phosphofructo-2-kinase/fructose-2,6-biphosphatase (PFKFB) promoter and HIF-1α 3′-untranslated region (UTR)] (14), and Dr. Shan Lu, University of Cincinnati (GAPDH promoter) (13). pGlut-1 wild type (Wt) and pGlut-1 mutant (Mt) were provided by Dr. Amit Maity, University of Pennsylvania (4); pGlut-1 (Wt) contains an active HRE, whereas the 4-bp mutation in pGlut-1 (Mt) disrupts HIF-1 binding. For transactivation studies of HIF-1, the binary Gal4 reporter plasmid (TAD aa 530–826) was provided by Dr. Nianli Sang (25). Backbone plasmid pM (Clonetech) contains no TAD but expresses the Gal4dbd. pFR-Luc (Stratagene) reporter contains the yeast Gal4-binding site upstream of a minimal promoter and the firefly luciferase gene. Construction of aggrecan luciferase promoter construct is described elsewhere (32). As an internal transfection control, vector pRL-TK (Promega) containing Renilla reniformis luciferase genes was used. In a previous study, we optimized the amount of transfected plasmid, the pretransfection period after seeding, and the posttransfection period before harvesting for rat nucleus pulposus cells using pSV β-galactosidase plasmid (Promega) (22).

Transfections and dual luciferase assay.

Nucleus pulposus and HeLa cells were cultured in 24-well plates at a density of 7.5 × 104 cells/well. Next day, cells were treated with 500 ng of reporter gene plasmid and 500 ng of control plasmid pTK-RL premixed with the transfection reagent Lipofectamine 2000 (Invitrogen, Carlsbad, CA). Twenty-four hours later, cells were transferred to the hypoxia chamber or maintained in normoxia for 1 additional day. The cells were then harvested, and a Dual-Luciferase reporter assay system (Promega) was used for sequential measurement of firefly and Renilla luciferase activities with specific substrates, i.e., beetle luciferin and coelenterazine, respectively. Quantification of luciferase activities and calculation of relative ratios were carried out using a luminometer (TD-20/20; Turner Designs, Sunnyvale, CA); at least three independent transfections were performed in triplicate.

Cell viability measurement by neutral red uptake.

Nucleus pulposus cells were treated with the following inhibitors: oligomycin (0.05–0.2 μM), antimycin A (10–30 μM), and atractyloside (50–200 μM) (Sigma). One day later, viability was measured using neutral red assay. Briefly, cells cultured in 24-well plates were treated with mitochondrial inhibitors in normoxia or hypoxia for 24 h. The cells were then incubated with neutral red solution dissolved in DMEM (40 μg/ml) for 3 h at 37°C. The dye solution was aspirated, and cells were washed with HBSS. The retained dye was extracted with 400 μl of acid ethanol (50% ethanol in 1% acetic acid) per well. The absorbance was read at 540 nm using a microplate reader (Tecan SpectraFluor Plus, Durham, NC). Results are presented as percent absorbance compared with untreated control values in normoxia.

ATP measurements following treatment with mitochondrial and glycolytic inhibitors.

Nucleus pulposus cells were cultured in six-well plates (2 ×105 cells/well) and treated with antimycin A (30 μM) or atractyloside (200 μM) for 24 h under normoxic (20% O2) conditions or cultured in hypoxia (2% O2). To examine the effects of inhibition of glycolysis on ATP production, cells were treated with 10 mM 2-deoxyglucose (2-DG; Sigma) for 30 min to 6 h. ATP measurement was performed using the ATP bioluminescence assay kit CLS II (Roche) per the manufacturer's protocol. Luminescence was measured at 562 nm using a luminometer (TD-20/20; Turner Designs). ATP levels were normalized to total protein, measured using the BCA assay (Pierce).

Mitochondrial fatty acid utilization by nucleus pulposus cells.

To learn whether nucleus pulposus cells possess functioning mitochondria, rat and human cells were plated in 12-well plates (2 ×105 cells/well) in complete growth medium for 24 h. Cells were then washed with PBS and suspended in preincubation medium (Krebs-Ringer bicarbonate medium containing 0.05% fatty acid-free BSA) or inhibitor medium (preincubation medium with 250 mg of glucose) for 1 h. After preincubation, the medium was removed and the cells were incubated in either 200 μl of myristate or 200 μl of palmitate incubation medium (110 μM fatty acid, 16.7 μCi/ml [3H]fatty acid, 0.05% BSA in PBS, pH 6.9). After 3 h, the medium was transferred to columns containing ∼2.9 ml of Dowex-1 ion exchange resin (no. 1X8–200; Sigma), which had been charged with 1.0 N NaOH and washed with Milli-Q water until the pH of the elute was the same as the wash water. The Dowex column serves to bind the unmetabolized fatty acids. The 3H-labeled water, produced by β-oxidation, was eluted with 2 ml of water from the Dowex. The activity of eluate was counted using a Beckmann LS 6500 liquid scintillation counter. Data were normalized to total protein measured using the Lowry method.

Measurement of aggrecan promoter activity following silencing of HIF-1α by Small Interfering RNA.

To learn whether HIF-1α activity regulates aggrecan expression, we silenced HIF-1α expression using small interfering (si)RNA technology. Briefly, nucleus pulposus cells were plated in 24-well plates and transfected with HIF-1α SMART Pool (Dharmacon) or control siRNA duplexes (Ambion) at a final concentration of 20 nM in DMEM using Lipofectamine 2000 (Invitrogen). Cells also received aggrecan or enolase-1 promoter constructs and pRL-TK plasmid at the time of transfection. Six hours after transfection, the medium was replaced with complete growth medium, and the cells were allowed to recover for 18 h. Cells were then cultured under normoxic or hypoxic conditions for an additional 24 h, and luciferase activity was assayed.

Statistical analysis.

Data are means (SD). To test for significance, data were analyzed using Student's t-test; the P values obtained are indicated in the text and legends. Each experiment was performed in triplicate.


HIF-1α protein expression and mRNA stability in nucleus pulposus cells is independent of the oxygen tension.

Figure 1A shows the relative expression levels of HIF-1α protein in rat nucleus pulposus cells under different culture conditions. At 2% O2 (hypoxia), the protein expression was similar to that measured at 20% O2 (normoxia). A similar level of expression was noted when the nucleus pulposus cells were treated with Dfx, a chemical agent used to mimic effects of hypoxia. Moreover, there was little change in HIF-1 immunofluorescence when maintained in hypoxia (Fig. 1B). To further examine the dependence of HIF-1α on oxygen tension, we analyzed mRNA expression using real-time RT-PCR. Figure 1C shows mRNA levels after exposure of nucleus pulposus cells to 2% O2 from 4 to 24 h. Compared with normoxia, no differences were apparent in HIF-1α transcript level in hypoxic nucleus pulposus cells for up to 24 h. We then analyzed 3′-UTR activity, as a measure of HIF-1α mRNA stability, in nucleus pulposus cells. We found that HIF-1α 3′-UTR activity was similar in both normoxic and hypoxic nucleus pulposus cells.

Fig. 1.

Expression of hypoxia-inducible factor (HIF)-1α protein and mRNA by nucleus pulposus (NP) cells under normoxic and hypoxic conditions. A: rat NP cells were cultured at 20% O2 (Nx) and 2% O2 (Hx); some cells were treated with desferrioxamine (Dfx), a hypoxia mimic, for 24 h. Nuclear proteins were separated and analyzed by Western blot using an antibody to HIF-1α. Multiple blots were quantified by densitometric analysis. Nucleus pulposus cells expressed HIF-1α in Nx. In Hx, there was a minimal increase in HIF-1α expression. Moreover, Dfx did not increase HIF-1α expression. B: NP cells were grown in Nx and Hx and treated with an antibody to HIF-1α. The cell nuclei were stained with propidium iodide. HIF-1α was expressed by NP cells in Nx. In Hx, there was a similar level of induction in HIF-1α in NP cells. Magnification, ×20. C: HIF-1α mRNA expression was measured using real-time RT-PCR in Nx and Hx. mRNA expression levels were similar, and there was no increase mRNA levels with increasing time in Hx. D: HIF-1α mRNA stability in Nx and Hx was studied by measuring the activity of the 3′-untranslated region (UTR). Luciferase activities in Hx and Nx were comparable. Values are means (SD) from 3 independent experiments. *P < 0.05. ns, Nonsignificant.

Nucleus pulposus cells show a modest increase in HIF-1α TAD activity in hypoxia.

We further assessed the activity of HIF-1α in nucleus pulposus cells by studying the activation of transactivation domain (TAD). Figure 2A shows that under hypoxic conditions, there was a modest but significant increase (40%, or 0.4-fold) in HIF-1 TAD activity in nucleus pulposus cells. HeLa cells, a model cell line widely used to evaluate the transcriptional responses of HIF-1α, was used as a control. In contrast to nucleus pulposus cells, HeLa cells displayed a significantly higher (3.5-fold) activation of HIF-1 TAD activity in hypoxia (Fig. 2B). These results indicate that hypoxia causes activation of TAD function in both nucleus pulposus and HeLa cells.

Fig. 2.

Oxemic regulation of HIF-1α-transactivation domain (TAD) transactivation activity in NP cells. NP (A) and HeLa cells (B) were transfected with a Gal4 binary reporter system consisting of Gal4-HIF1-TAD and pFR-Luc reporter. Gal4dbd (empty vector pM) was used to measure background activity. There was a 40% increase (0.4-fold) in TAD activity when nucleus pulposus cells were cultured in hypoxic (2% O2) conditions. In contrast, HeLa cells showed a much higher (3.5-fold) induction in TAD activity at 2% O2. Values are means (SD) of 3 independent experiments. *P < 0.05.

Oxemic regulation of HIF-1α target genes by nucleus pulposus cells.

Real-time RT-PCR analysis showed that expression of GLUT-1, GLUT-3, and GAPDH were unaffected by the oxemic status of the nucleus pulposus cells (Fig. 3, AC). In terms of copy number relative to β-actin, expression of all these genes was similar under normoxic and hypoxic conditions. We also evaluated GAPDH protein levels in the nucleus pulposus cells. Figure 2D indicates that culture time in hypoxia did not change the level of expression of this glycolytic protein. Moreover, in hypoxia, Western blot followed by densitometric analysis showed that expression levels of GLUT-1 and GLUT-3 in the membrane, as well as total protein fraction, was independent of time or the oxygen tension (Fig. 3, DF). Similarly, immunofluorescence analysis of normoxic and hypoxic nucleus pulposus cells indicated that there was a comparable level of staining of GLUT-1 (Fig. 3G).

Fig. 3.

Oxemic regulation of HIF-1α target genes by NP cells. Cells were cultured in Nx or Hx for 4–24 h. AC: real-time RT-PCR was performed to study the relative expression levels of GLUT-1 (A), GLUT-3 (B), and GAPDH (C). In Hx, there was no increase in mRNA levels of GLUT-1, GLUT-3, and GAPDH. There was no time-dependent increase in mRNA expression of all 3 genes. Values are means ± SE from 3 independent experiments. D: expression levels of GLUT-1, GLUT-3, and GAPDH in total extract and the membrane protein fraction resolved by Western blot analysis. Blots were probed with antibodies against GLUT-1, GLUT-3, or GAPDH. GAPDH expression levels in Nx were also comparable to its levels in Hx. β-Actin was used as a control and shows equal loading. E and F: multiple Western blots of total GLUT-1 (E) and GLUT-3 protein (F) were quantified by densitometric analysis. In both the membrane and total cell extract, expression levels of GLUT-1 and GLUT-3 were similar in Hx and Nx and were not dependent on time. G: detection of GLUT-1 expression by immunofluorescence microscopy. After culture in Hx or Nx for 24 h, cells were fixed and stained with antibody against GLUT-1. In Nx and Hx, cells showed similar levels and patterns of GLUT-1 staining.

To further evaluate the transcriptional regulation of HIF-1α target genes, we examined the promoter activities of GLUT-1 and GLUT-3 under normoxic and hypoxic conditions. For these studies, we used luciferase reporter constructs containing wild-type or mutant HRE (21). HeLa cells were used as a control. Figure 4, A and B, shows that promoter activities of GLUT-1 and GLUT-3 in nucleus pulposus cells were similar in both hypoxia and normoxia. In contrast, in hypoxia, HeLa cells evidenced a high induction of GLUT-1 (125-fold) and GLUT-3 (15-fold) promoter activities (Fig. 4, C and D). Furthermore, in nucleus pulposus cells, basal normoxic promoter activity of GLUT-1 was two- to threefold higher than background activity measured with GLUT-1 control plasmid containing mutant HRE. The normoxic basal promoter activity of HeLa cells was minimal and similar to the background activity measured with the HRE mutant GLUT-1 plasmid. Moreover, when transfected with a mutated control GLUT-1 plasmid, there was little difference between normoxic and hypoxic luciferase activities of both nucleus pulposus (Fig. 4A) and HeLa cells (Fig. 4C).

Fig. 4.

Evaluation of promoter activities of HIF-1α target genes in NP and HeLa cells cultured in Nx and Hx. GLUT-1 (A and C) or GLUT-3 (B and D) reporter plasmids containing a wild-type (Wt) or mutant (Mt) hypoxia-response element (HRE), driving firefly luciferase, were transfected into rat nucleus pulposus cells along with the pRL-TK vector. Cells were cultured in Nx (20% O2) or Hx (2% O2) for 24 h, and luciferase reporter activity was measured. When the oxygen tension was decreased to 2%, there was no induction of GLUT-1 (A) or GLUT-3 reporter activity (B) in NP cells. In contrast, in Hx, HeLa cells evidenced 120-fold induction of GLUT-1 (C) and 15-fold induction of GLUT-3 promoter activities (D). Cells transfected with Mt-GLUT-1 plasmid evidenced very low luciferase activity. NP cell Wt-GLUT-1 reporter activity was ∼2.5-fold higher than background activity, measured using Mt-GLUT-1 plasmid, indicating the presence of transcriptionally active HIF-1α. In contrast, in HeLa cells, normoxic activity of Wt-GLUT-1 reporter was similar to background activity. Values are means (SD) of 3 independent experiments performed in triplicate (n = 3). *P < 0.05.

In addition, we analyzed promoter activities of two key glycolytic enzymes, GAPDH and PFKFB (Fig. 5, AD). In accordance with GLUT-1 and GLUT-3 promoter activity, GAPDH promoter activity showed no hypoxic induction (Fig. 5A); in contrast, HeLa cells displayed a 50- to 60-fold induction of GAPDH promoter activity in hypoxia (Fig. 5C). However, unlike GLUT and GAPDH, there was a threefold induction of PFKFB promoter activity in hypoxia in nucleus pulposus cells (Fig. 5B). Hypoxic induction of PFKFB promoter activity in HeLa cells was close to 120-fold (Fig. 5D). In addition, real-time PCR analysis indicated that there was a significant increase in PFKFB mRNA expression in nucleus pulposus cells cultured in hypoxia (Fig. 5E).

Fig. 5.

A and B: evaluation of promoter activities of HIF-1α target genes in NP and HeLa cells cultured in Nx and Hx. GAPDH (A and C) or 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase (PFKFB) reporter plasmids (B and D) containing a Wt HRE, driving firefly luciferase, were transfected into rat nucleus pulposus cells along with pRL-TK vector containing the Renilla luciferase gene. Cells were cultured in Nx (20% O2) or Hx (2% O2) for 24 h, and luciferase reporter activity was measured. When the oxygen tension was decreased to 2%, there was no induction of GAPDH reporter activity (A) in NP cells. In contrast, in Hx, PFKFB promoter showed a 3-fold induction in NP cells (B). Promoter activities of GAPDH (C) as well as PFKFB (D) in HeLa cells evidenced a 55- to 120-fold induction, respectively, in Hx. Values are means (SD) of 3 independent experiments performed in triplicate (n = 3). *P < 0.05. E: real-time RT-PCR was performed to study the relative expression levels of PFKFB in normoxic and hypoxic NP cells. In Hx, there was a significant increase in mRNA levels of PFKFB. Values are means (SD) of 3 independent experiments. *P < 0.05.

Nucleus pulposus cells generate energy through glycolysis.

To investigate the mode of energy generation in nucleus pulposus cells, we measured ATP production. Under basal normoxic conditions, ATP levels in nucleus pulposus cells were between 20 and 25 nM/mg protein (Fig. 6A). When treated with 2-DG, an inhibitor of glycolysis, basal ATP levels are suppressed by almost 80% within 30 min of treatment and remained very low for 6 h (Fig. 6A). In contrast, when cells were treated with the mitochondrial function inhibitors antimycin A or atractyloside, there was no effect on basal ATP levels after 30 min or 6 h of treatment (Fig. 6B). We also evaluated cell viability after treatment with mitochondrial inhibitors in normoxia or hypoxia using the neutral red uptake assay. Figure 6C indicates that cell viability in both normoxia and hypoxia (data not shown) was unaffected.

Fig. 6.

Energy metabolism in NP cells studied using inhibitors of glycolysis and oxidative phosphorylation. A: nucleus pulposus cells in Nx were treated with 2-deoxyglucose (2-DOG; 10 mM), and ATP yield was measured. Basal ATP levels in NP cells in Nx ranged from 20 to 25 nM/mg protein, which was higher in Hx. Treatment with 2-DOG resulted in significant (75–80%) suppression of ATP yield within 30 min of treatment; levels remained suppressed for 6 h. B: NP cells were treated with increasing concentrations of the mitochondrial function inhibitors antimycin A (10 and 30 μM) or atractyloside (50 and 200 μM) for 6 h. There was little change in ATP yield after treatment with both inhibitors. C: cell viability after treatment with the mitochondrial inhibitors oligomycin (0.2 μM), antimycin A (30 μM), and atractyloside (200 μM) in normoxia was evaluated using the neutral red uptake assay. There was no decrease in cell viability after treatment with any of the inhibitors or when cultured in Hx. Values are means (SD) of 3 independent experiments performed in triplicate (n = 3). *P < 0.05.

Nucleus pulposus cells possess functioning mitochondria.

We determined that nucleus pulposus cells contained functioning mitochondria by measuring the oxidation of free fatty acids by the β-oxidation pathway. Mitochondria of both human and rat nucleus pulposus cells oxidized [3H]palmitate (C16:O) and myristate (C14:O) into CO2 and water (Fig. 7, AC). For both human and rat nucleus pulposus cells, rates of oxidation of palmitate were comparable (Fig. 7A). Myristate was oxidized at a significantly faster rate by the rat cells compared with human nucleus pulposus cells. When excess glucose was included in the incubation medium along with 3H-labeled palmitate (Fig. 7B) and myristate (Fig. 7C), there was a significant inhibition (90–95%) of lipid oxidation and a concomitant decrease in [3H]H2O generation.

Fig. 7.

Free fatty acid oxidation by mitochondria of NP cells. A: NP cells were incubated with 3H-labeled myristate or palmitate, and the released [3H]H2O was measured by scintillation counting. Rate of fatty oxidation was calculated as oxidation of pmol fatty acid·min·mg total protein−1. A: both human (Hu) and rat (Rt) NP cells oxidized palmitate and myristate as an energy source. Human and rat NP cells oxidized palmitate at similar rates; myristate was metabolized faster by the rat NP cells. B and C: utilization of both fatty acids by NP cells was blocked by inclusion of glucose (Glc) in the incubation medium. In the presence of Glc, there was a significant suppression (90–95%) in [3H]H2O liberated from both palmitate and myristate for human and rat cells. Values are means (SD) of 3 experiments. *P < 0.05.

HIF-1α expression and glycolysis is required for maintaining aggrecan promoter activity in nucleus pulposus cells.

The functional significance of constitutive HIF-1α expression and glycolytic metabolism in nucleus pulposus cells was studied by measuring aggrecan expression and gene promoter activity. We suppressed HIF-1α expression and function or treated cells with 2-DG. Figure 8A shows that in normoxia, when cells were cotransfected with HIF-1α SiRNA, aggrecan promoter activity was inhibited by almost 60%. Similarly, when HIF-1α was silenced under hypoxic conditions, there was a comparable level of aggrecan promoter suppression. We confirmed that HIF-1α was silenced by determining PFKFB promoter activity in normoxic and hypoxic culture conditions. Figure 8B indicates that the silenced cells exhibited a significant decrease in the activity of the PFKFB gene promoter, a known HIF-1 target gene. Similarly, treatment with 17-AAG, a known HIF-1α inhibitor, resulted in a significant drop in aggrecan mRNA expression (Fig. 8C). In the presence of the glycolysis inhibitor 2-DG, there was ∼30% suppression of aggrecan promoter activity under normoxic conditions (Fig. 8D).

Fig. 8.

HIF-1α regulation of aggrecan expression in NP cells. Cells were cotransfected with aggrecan promoter construct (A) or PFKFB reporter (B) with either HIF-1α-siRNA (si-HIF-1) or control-SiRNA duplexes (Co). The transfected cells were cultured under Nx or Hx for 24 h. Silencing of HIF-1α in Nx or Hx evidenced a significant decrease in aggrecan promoter as well PFKFB promoter activity. C: NP cells were treated with 17-allylamino-17-demethoxygeldanamycin (17-AAG), a HIF-1α inhibitor, and aggrecan gene expression was assessed using real-time RT-PCR analysis. Suppression of HIF-1a activity resulted in decreased expression of aggrecan expression in both Nx and Hx. D: NP cells were treated with 2-DOG (10 mM) for 24 h in Nx, and aggrecan promoter activity was measured. The inhibition of glycolysis caused 30% inhibition of promoter activity. Values are means (SD) of 3 independent experiments. *P < 0.05.


Previously, we showed that the HIF-1α isoform is expressed in the nucleus pulposus in situ, as well as in cells cultured in 20% O2 (20, 22). Since a major role of HIF-1α is to maintain cell function at a low Po2, we initially explored the mechanism by which the oxemic state regulated the expression of this critical transcription factor. Predictably, Western blot and immunofluorescence analysis showed that at 2% O2, HIF-1α was robustly expressed by nucleus pulposus cells. Surprisingly, when cells were maintained at 21% O2, there was little change in HIF-1α expression. Likewise, chemical hypoxia did not significantly increase HIF-1α expression, and when HIF-1α mRNA stability was evaluated, quantitative PCR failed to indicate differences in steady-state transcript levels in normoxia and hypoxia. Thus, from all perspectives, HIF-1α expression in nucleus pulposus cells was found to be independent of the oxygen tension. This unique finding points to substantive underlying differences between the metabolism of disk cells and cells from most other tissues. Importantly, since HIF-1α serves a critical role in regulating a host of other (nonmetabolic) target genes, it is likely that this transcription factor influences a number of other survival activities (2). In the current study, we have shown that HIF-1α is required for the synthesis of the water-binding proteoglycan molecules of the extracellular matrix.

One obvious question raised by the observation that HIF-1α was expressed in 20% O2 concerned the activity of target genes: did HIF-1α transactivate genes of the glycolytic pathway? Thus, although the transcription factor was robustly expressed by the nucleus pulposus, it was not known whether expression enhanced target gene activity. Focusing on those genes that regulate the energetic state of nucleus pulposus cells, we assessed expression of the glucose transporters GLUT-1 and GLUT-3. In the same study, we also evaluated another HIF-1 target, GAPDH, the enzyme that catalyzes the conversion of glyceraldehyde 3-phosphate into 1,3-bisphosphoglycerate (13). At 2 and 20% O2, GLUT-1, GLUT-3, and GAPDH mRNA expression levels were comparable and remained constant over time. Furthermore, at the two disparate oxygen tensions, both Western blot and immunofluorescence analysis indicated comparable GLUT-1, GLUT-3, and GAPDH protein levels. In addition, GLUT-1 and GLUT-3 promoter activities were not responsive to changes in oxygen tension, and there were no oxygen-dependent changes in GAPDH promoter activity. It is relevant to note that although these genes were not responsive to the oxemic state of the culture, we did observe a threefold induction in PFKFB promoter activity; the magnitude of the change was similar to that observed when enolase-1 promoter activity was measured in hypoxic nucleus pulposus cells (22). For these two genes, hypoxic induction of glycolytic promoter activity was probably due to a modest increase in HIF-1α transactivation activity. For comparison, our studies with HeLa cells, which are clearly HIF responsive, indicate that these modest changes in transactivation activity would not substantially increase the rate of glycolysis by nucleus pulposus cells. In the HIF-1α-responsive HeLa cell, the change from normoxia to hypoxia caused a 60- to 120-fold increase in promoter activities and a 3.5-fold elevation in TAD activity. All of these values are substantially different from those exhibited by nucleus pulposus cells. On the basis of this observation, it is concluded that the oxemic stability of the nucleus pulposus cells is optimal for survival in an environment where there are frequent shifts in vascular supply and oxygen delivery; in the intervertebral disk, these shifts may reflect minute-to-minute variations in applied biomechanical forces to the spinal units.

Since resistance to compressive forces on the spine are mediated by water molecules bound to charged sites on proteoglycans, we examined the relationship between HIF-1α expression and aggrecan promoter activity. Previous studies by Pfander et al. (18) showed that under hypoxic conditions, aggrecan mRNA and sulfation levels were significantly reduced in chondrocytes lacking HIF-1α. In a more recent study, Robins et al. (23) reported that hypoxia via a HIF-1α-dependent mechanism increased the commitment of mesenchymal cells to the chondrogenic lineage by activating Sox-9; one important aspect of commitment is the promotion of aggrecan synthesis. Studies reported presently build on these earlier observations. We have shown that irrespective of the oxemic state of the cells, when HIF-1α expression was blocked by siRNA or by pharmacological agents, aggrecan promoter activity was markedly inhibited. A similar level of inhibition was observed when glycolysis was blocked. On the basis of these observations, we suggest that HIF-1α promotes aggrecan expression directly, by elevating gene expression and protein synthesis, and, possibly, indirectly, through lineage specification as well as promotion of sulfation reactions. From a functional viewpoint, since aggrecan synthesis is dependent on HIF-1α expression, rather than oxemic status, long-term shifts in the oxemic state of the disk would be expected to exert a minimal effects on the generation of extracellular matrix components.

One of the logical consequences of stabilization of HIF-1α and the robust expression of glucose transporters and glycolytic enzymes is upregulation of anaerobic glycolysis. We found that in normoxia, basal concentrations of ATP in nucleus pulposus cells were between 20 and 25 nM/mg protein. These values are comparable to levels reported for articular chondrocytes (18), another cell type that uses glycolysis to generate energy (29, 30). In the presence of 2-DG, a potent inhibitor of glycolysis, ATP generation was suppressed almost 80%. The sensitivity of the cells to this inhibitor emphasized the reliance on glycolysis for energy generation. Together, these data confirm earlier findings that the cells of the intervertebral disk generate much, if not all, of their energy through anaerobic glycolysis (1, 10).

Although ATP generation through glycolysis is clearly the major pathway for energy generation, the possibility exists that some ATP may be generated through mitochondrial oxidative phosphorylation. We noted that when cells were treated with inhibitors of mitochondrial function, no change in ATP production, or viability, was observed. Indeed, this observation raised the question, are there functioning mitochondria in cells of the nucleus pulposus? Although, this topic has received minimal attention, Gan et al. (7) reported earlier that although nucleus pulposus cells contained mitochondria with normal architecture, the total number of organelles per cell was very low. To further address this issue, we examined fatty acid oxidation in the absence of a three-carbon source. Since cells from both human and rat tissues oxidized palmitate (23–27 pmol·min−1·mg protein−1) and myristate (25–43 pmol·min−1·mg protein−1), it was clear that the nucleus pulposus had the capacity to generate ATP through oxidative metabolism. Moreover, we showed that in concert with other tissues, if nucleus pulposus cells were provided with excess glucose, fatty acid metabolism was suppressed. These results lends strong support to the notion that although glucose and anaerobic glycolysis represent the major fuel and pathway for energy generation, respectively, the nucleus pulposus retains the capacity to metabolize fatty acids through mitochondrial oxidative metabolism.

Finally, there is now good evidence to indicate that HIF-1α plays a major role in directing the interplay between glycolysis and mitochondrial oxidative phosphorylation (11, 17). These workers have shown that HIF-1α inhibits mitochondrial function by trans-activating the gene encoding pyruvate dehydrogenase kinase 1. Since this protein suppresses pyruvate dehydrogenase, pyruvate cannot be converted into acetyl-CoA, and as a result, the TCA cycle is blocked. On the basis of these observations, it is concluded that although mitochondrial function is retained by cells of the intervertebral disk, it is reasonable to assume that normoxic expression of HIF-1α by nucleus pulposus cells serves to suppress oxidative phosphorylation, lower mitochondrial ATP generation, and promote glycolytic ATP and aggrecan synthesis. These two latter activities are of considerable functional importance and point to the role of HIF-1α in promoting the synthesis of key water-binding molecules of the extracellular matrix, complexes that are required to maintain the biomechanical properties of the intervertebral disk.


This work was supported by National Institute of Arthritis and Musculoskeletal and Skin Diseases Grant AR050087.


We thank Drs. Sherin Devaskar, Oleksandr Minchenko, Shan Lu, Amit Maity, and Nianli Sang for providing necessary plasmids.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract