Endocrinology and Metabolism

Dexamethasone-induced expression of the glucocorticoid response gene lipocalin 2 in chondrocytes

H. C. Owen, S. J. Roberts, S. F. Ahmed, C. Farquharson


Glucocorticoids (GC) are commonly used anti-inflammatory drugs, but long-term use can result in marked growth retardation in children due to their actions on growth plate chondrocytes. To gain an insight into the mechanisms involved in GC-induced growth retardation, we performed Affymetrix microarray analysis of the murine chondrogenic cell line ATDC5, incubated with 10−6 M dexamethasone (Dex) for 24 h. Downregulated genes included secreted frizzled-related protein and IGF-I, and upregulated genes included serum/GC-regulated kinase, connective-tissue growth factor, and lipocalin 2. Lipocalin 2 expression increased 40-fold after 24-h Dex treatment. Expression increased further after 48-h (75-fold) and 96-h (84-fold) Dex treatment, and this response was Dex concentration dependent. Lipocalin 2 was immunolocalized to both proliferating and hypertrophic growth plate zones, and its expression was increased by Dex in primary chondrocytes at 6 h (3-fold, P < 0.05). The lipocalin 2 response was blocked by the GC-receptor antagonist RU-486 and was increased further by the protein synthesis blocker cycloheximide. Proliferation in lipocalin 2-overexpressing cells was less than in control cells (49%, P < 0.05), and overexpression caused an increase in collagen type X expression (4-fold, P < 0.05). The effects of lipocalin 2 overexpression on chondrocyte proliferation (64%, P < 0.05) and collagen type X expression (8-fold, P < 0.05) were further exacerbated with the addition of 10−6 M Dex. This synergistic effect may be explained by a further increase in lipocalin 2 expression with Dex treatment of transfected cells (45%, P < 0.05). These results suggest that lipocalin 2 may mediate Dex effects on chondrocytes and provides a potential novel mechanism for GC-induced growth retardation.

  • microarray
  • growth retardation

glucocorticoids (GC) are commonly used as anti-inflammatory or immunosuppressive treatments, and it has been estimated that 5–10% of children may require some form of GC therapy during childhood (67). Since the introduction of GCs in the treatment of rheumatoid arthritis over 50 years ago, the therapeutic applications of these drugs have greatly broadened to encompass a large number of non-endocrine and endocrine diseases (28). Despite intense efforts in optimizing therapy, adverse effects of GC medication are still common, with growth impairment being one of the commonest in childhood (48, 49).

Normal cartilage formation and bone growth are dependent on chondrocyte progression through the growth plate, as demonstrated by a number of null mutations that have been introduced in mice. The rate of longitudinal bone growth is controlled through the precise regulation of chondrocyte proliferation, differentiation, hypertrophy, and apoptosis (18, 36). The mechanisms by which GCs induce failure in the growth process are presently unclear but are likely to include antagonism of growth hormone (GH) secretion and action. It has been shown that GCs inhibit pulsatile GH release (26, 27, 68), reduce GH receptor expression, and inhibit IGF-I activity (64). GCs also act locally to inhibit longitudinal bone growth, which suggests a mechanism intrinsic to the growth plate. Baron and colleagues (3) reported that local infusion of GC into the growth plate decreased tibial growth with no effects on the contralateral limb. In addition, both Silvestrini et al. (59) and Smink et al. (60) reported that GC reduced growth cartilage width by inhibiting chondrocyte proliferation and increasing the apoptosis of terminal hypertrophic chondrocytes. Studies by our group and others have also shown that GCs impair chondrocyte proliferation and increase apoptosis (33, 35, 50, 53, 55, 61). However, the mechanisms by which GCs exert control over chondrocyte dynamics are presently unclear. It is likely that GCs affect a wide range of complex regulatory and signaling networks involving cell matrix and intercellular interactions to mediate chondrocyte proliferation and differentiation.

In this study, Affymetrix microarray gene expression profiling of ATDC5 cells was used to identify novel GC-responsive chondrocyte genes in an attempt to further our understanding of the processes involved in GC-induced growth retardation. We then proceeded to characterize the regulation, expression, and function of lipocalin 2, a secreted acute-phase protein that is thought to have an anti-inflammatory function and is involved in the transport of small hydrophobic molecules (21). Lipocalin 2-deficient mice have compromised immunity when subjected to bacterial infections but have no other obvious phenotype (5, 22). In this study, we show that lipocalin 2 is markedly increased in response to dexamethasone (Dex) in chondrocytes and may play a role in exacerbating GC effects on the growth plate.


Cell culture.

The ATDC5 chondrocyte cell line was obtained from the RIKEN Cell Bank (Ibaraki, Japan), and cells were cultured at a density of 6,000 cells/cm2 in differentiation medium (DMEM-F12; Invitrogen, Paisley, UK), 5% FCS (Invitrogen), 3 × 10−8 M sodium selenite, 10 μg/ml human transferrin (Sigma, Poole, UK), and 10 μg/ml insulin (Sigma). Cells were differentiated at 37°C in a humidified atmosphere containing 5% CO2 for 8–15 days, and Dex (final concentration 10−6 M) was added to the cells at the indicated times. We and others have previously shown that, from 8 days of differentiation, ATDC5 cells express the chondrocyte marker gene collagen type II, and by day 15 the cells display a terminally differentiating phenotype with increased levels of collagen type X and aggrecan expression and the formation of nodules (49, 16). To test lipocalin 2 expression following exposure to other antiproliferative treatments, ATDC5 cells were cultured in the presence of ceramide (40 μ M), TNFα (10 ng/ml), or serum-free medium (SFM) for 48 h. The protein synthesis blocker cycloheximide (CHX; 8 μg/ml), and the GC receptor antagonist RU-486 (10−5 M) were added to ATDC5 cells to investigate the mechanisms involved in Dex-induced lipocalin 2 expression. Murine recombinant lipocalin 2 was added to the cells at a range of concentrations (5–20 μg/ml; R&D Systems, Abingdon, UK) to analyze the effect of lipocalin 2 on cell proliferation. Control cells received vehicle only.

cDNA microarray hybridization.

Total RNA was extracted from duplicate control and Dex-treated cultures at day 15 of differentiation 24 h following treatment using previously described methods (52). Briefly, total RNA was extracted from chondrocytes by repeated aspiration though a 25-gauge syringe needle in 1 ml of Ultraspec (Biotecx, Houston, TX). After extraction with chloroform, RNA in the aqueous phase was purified with isopropanol and bound to 50 μl of RNA Tack resin before washing with 75% ethanol. The RNA was then eluted in 100 μl of nuclease-free water and treated with DNase. The quality of each sample was considered suitable by a 260/280 ratio of greater than 1.9. Samples were sent to the UK Human Genome Mapping Programme-Resource Centre (Cambridge, UK), where target preparation and hybridization to the Affymetrix Murine 430 2.0 GeneChip was carried out following the standard Affymetrix protocol (http://www.affymetrix.com/support/technical/manuals.affx). This GeneChip contains ∼39,000 full-length mouse genes and EST clusters from the UniGene database. The data obtained from these hybridizations can be viewed on the EMBL-EBI Array Express repository (http://www.ebi.ac.uk/microarray-as/aer/?#ae-main[0]) (accession no. E-MEXP-1244).

Microarray data analysis.

The data obtained from the Affymetrix hybridization was preprocessed using the RMA algorithm in GeneSpring 7.0 (Silicon Genetics). Each gene was then normalized to the control sample. To assess differential gene expression between treatments, expression values were further filtered by retaining only those probe sets with a fold change of at least 1.5 in Dex samples compared with Controls. A t-test was then carried out, and genes associated with a P value < 0.05 were regarded as statistically significant. This analysis resulted in a gene list of 96 transcripts, whose expression was significantly changed 1.5-fold or more in Dex-treated samples. Probe set lists resulting from the comparison of Control vs. Dex samples filtered using a 1.5-fold cut-off were assigned a molecular function with the NetAffx annotation programme (http://www.affymetrix.com/analysis). The choice of candidates from our short list that were initially prioritized for future study was based on reviews of function, likely relationship to GC action, and association with chondrocytes and bone growth.

Quantitative PCR.

Quantitative PCR (qPCR) was used to confirm changes in the expression of selected genes. ATDC5 cells at day 15 of culture were incubated with 10−6 M Dex for 24 h and RNA was extracted. RNA samples or blanks (containing nuclease-free water in place of RNA) were reverse transcribed in 20-μl reactions with 200 ng of random hexamers and 200 U of Superscript II reverse transcriptase using the Superscript preamplification protocol (Invitrogen). qPCR was performed using the Stratagene Mx3000P Real-Time QPCR System (Stratagene, La Jolla, CA). Primers were designed using the software program Primer3 (Whitehead Institute for Biomedical Research) and were made to span at least one intron to identify any amplification from contaminating genomic DNA by semiquantitative PCR. The primers were as follows: lipocalin 2, forward 5°ä-CAGAAGGCAGCTTTACGATG-3′, reverse 5′-CCTGGAGCTTGGAACAAATG-3′, product 134 bp; secreted frizzled-related protein-2, forward 5′-TACCACGGAAGCCTCTAAGC-3′, reverse 5′-CTCGCTTGCACAGAGATGTT-3′, product 100 bp; connective tissue growth factor (CTGF), forward 5′-CCACCCGAGTTACCAATGAC-3′, reverse 5′-GACAGGCTTGGCGATTTTAG-3′, product 146 bp; IGF-I, forward 5′-GTGGACCGAGGGGCTTTTACTT-3′, reverse 5′-TTTGCAGCTTCGTTTTCTTGTTTG-3′, product 246 bp (from 7) Lumican, forward 5′-TGCTCGAGCTTGATCTCTCC-3′, reverse 5′-CAGTGGTCCCAGGATCTTACA-3′, product 156 bp; integrin α10, forward 5′-CTGAGGCTGGTTCACAATGA-3′, reverse 5′-CGGGAGGCTTCATTCAGTAG-3′, product 138 bp; dentin matrix protein-1 (DMP1), forward 5′-AAAGTCAAGCTAGCCCAGAGG-3′, reverse 5′-CCGGTCCCCGTACTCTTAG-3′, product 129 bp; serum glucocorticoid-regulated kinase (SGCK), forward 5′-GATGGGCCTGAACGATTTTA-3′, reverse 5′-GAGGAGAGGGGTTAGCGTTC-3′, product 111 bp. cDNA (10 ng) derived from three control and three Dex-treated cultures were amplified in triplicate using the Platinum SYBR Green qPCR SuperMix (Invitrogen) under the following conditions. cDNA was denatured for 2 min at 95°C, followed by 40 cycles consisting of 15 s at 95°C and 30 s at 60°C and 1 cycle consisting of 1 min at 95°C and 30 s at 60°C. Each reaction contained 10 μl (10 ng) of cDNA, 25 μl of Platinum SYBR Green qPCR Super-Mix, 1 μl of ROX Reference Dye, and 250 nM primer in a total volume of 50 μl. A dilution series of both gene of interest and external control were carried out and subjected to an identical PCR to allow estimation of PCR efficiency, and fold changes normalized for the expression of GAPDH were calculated using the comparative method (45). Initial experiments confirmed that lipocalin 2 expression was increased by a similar level during both chondrogenesis (day 8) and terminal differentiation (day 15) (data not shown). Consequently, for characterization of lipocalin 2 expression in ATDC5 cells, Dex (10−6 to 10−9 M) was added to the cells on day 8 of culture for the stated time periods.

Western blotting.

Cells were lysed in RIPA buffer (20m M Tris·HCl, 135 mM NaCl, 10% glycerol, 1% Igepal, 0.1% SDS, 0.5% deoxycholic acid, 2 mM EDTA) containing protease inhibitors (Complete EDTA-free protease inhibitor cocktail; Roche, Mannheim, Germany), and protein content was quantified using the DC-protein assay reagent (Bio-Rad, Hercules, CA). Protein lysates (15 μg) were loaded onto 10% bis-tris gels, fractionated, and electroblotted onto nitrocellulose membranes. After blocking with 5% nonfat milk powder and 0.1% Tween, the membranes were incubated overnight with anti-lipocalin 2 primary antibody (1:500; R&D Systems) at 4°C, washed, and incubated at room temperature for 1 h with rabbit anti-goat peroxidase-labeled secondary antibody (1:2,000; DAKO, Cambridge, UK). β-Actin expression was also measured as a loading control (anti-β-actin clone AC15, Sigma A5441, 1:5,000). The immune complexes were then visualized by enhanced chemiluminescence (GE Healthcare, Buckinghamshire, UK).

Analysis of ATDC5 proliferation with antiproliferative treatments.

Chondrocyte proliferation was assessed by incubating SFM, TNFα (10 ng/ml), or ceramide-treated (40 μM) ATDC5 cells at day 8 of maturation with 0.2 μCi/ml [3H]thymidine (37 MBq/ml; Amersham Pharmacia Biotech, Buckinghamshire, UK) for the last 2 h of the incubation period. Following removal of [3H]thymidine, the cells were fixed in ice-cold TCA (5%) for 15 min, washed, and lysed in 0.1 M NaOH for up to 30 min. Scintillation fluid was then added, and the amount of radioactivity incorporated into TCA-insoluble precipitates was measured using a scintillation counter.

Histological analysis of lipocalin 2 expression in the growth plate.

Tibiae were dissected from 10-day-old and 5-wk-old C57BL6xCBA mice that had been euthanized by cervical dislocation. Metatarsals were dissected from 17-day-old fetal Swiss mice euthanized by cervical dislocation. All mice were maintained by the Roslin Institute Small Animal Unit in accordance with Home Office guidelines for the care and use of laboratory animals. The limbs were fixed overnight in 4% paraformaldehyde at 4°C, decalcified for 5 days in 10% EDTA (pH 8.0), dehydrated, and embedded in paraffin before sectioning at 6 μm. For histological analysis, sections were dewaxed and demasked for 30 min in 10 mM citrate buffer. Endogenous peroxidases and nonspecific antibody binding were blocked with 3% H2O2 and normal goat serum, respectively, before overnight incubation at 4°C with 2 μg/ml anti-lipocalin 2 (R&D Systems). Following this, the sections were washed in PBS and incubated with rabbit anti-goat IgG peroxidase (1:100 dilution; DAKO) for 60 min. Diaminobenzidine (DAB) substrate reagent (0.06% DAB, 0.1% H2O2 in PBS) was incubated for 8 min, rinsed in PBS, and counterstained with hematoxylin (Zymed, Invitrogen) for 5 min. The sections were then dehydrated and mounted in DPX.

Isolation of primary murine chondrocytes.

Primary cultures of chondrocytes from rib growth plates of newborn Swiss mice were prepared using the isolation procedure and culture system developed by Lefebvre et al. (41). Briefly, the rib cage and sternum were dissected from 1-day-old mice, rinsed in PBS, and incubated at 37°C for 30 min in 2 mg/ml pronase in PBS and then in 3 mg/ml collagenase D in DMEM for 2 h. Soft tissues were detached from the cartilage by repeated pipetting. The denser cartilage was allowed to sediment, and the soft tissues were aspirated; the cartilage was then further digested with collagenase D for 3 h. Undigested bony parts were discarded, and the remaining cell suspension was aspirated repeatedly with a pasteur pipette, filtered through a 45-μm cell sieve, rinsed first in PBS and then in serum-free DMEM, and cells were counted. Chondrocytes were seeded in six-well plates at a density of 100,000 cells per well in DMEM with antibiotics, ascorbic acid, and 10% FBS. The cells were then cultured for 24 h, at which point Dex (final concentration 10−6 M) was added to the cells for 6, 24, or 72 h.

Generation of lipocalin 2-overexpressing ATDC5 cells.

Lipocalin 2 was amplified from ATDC5 cDNA and inserted into the EcoRI site of the pWGB10 overexpression vector (a kind gift from Dr. Wei Cui, Imperial College, London, UK). One microgram of pWGB10-lipocalin 2 was transfected into 80% confluent ATDC5 cells with FuGene6 according to the manufacturer's instructions (Roche Diagnostics, Basel, Switzerland). Control cell lines were generated using the same procedure with empty pWGB10 vector. After 48 h, cells were selected in medium containing 2 μg/ml puromycin, and after 2 wk clones were isolated and lipocalin 2 expression was evaluated by PCR and Western blotting.

Effect of lipocalin 2 overexpression on chondrocyte proliferation and differentiation.

To understand more fully the functional effects of lipocalin 2 overexpression on chondrocyte function, we assessed chondrocyte proliferation ([3H]thymidine uptake) and chondrocyte differentiation (collagen type X expression). Chondrocyte proliferation was assessed in transfected ATDC5 cells at day 8 of maturation, as previously described (49; also detailed above). Collagen type X expression was measured by qPCR as described above. These experiments determined whether the response of ATDC5 cells to Dex treatment (10−6 M for 48 h) was modified in the presence of lipocalin 2 overexpression.

Statistical analysis.

Analysis of variance (ANOVA) was performed to determine statistical significance. General linear model analysis incorporating pairwise comparisons using Tukey's test was used to compare groups within the ANOVA models. All data are expressed as means ± SE. Statistical analysis was performed using Minitab 14. Statistical significance was accepted at P < 0.05.


Identification and validation of dex-responsive chondrocyte genes.

Total RNA isolated from chondrogenic ATDC5 cells was hybridized to the Affymetrix Murine 430 2.0 Array. Analysis in GeneSpring 6.1 software demonstrated that from 39,000 transcripts the expressions of 96 genes were significantly changed 1.5-fold or more (Tables 1 and 2). Gene ontology analysis revealed that a large proportion of the genes were involved in cell death, proliferation, skeletal development, cell signaling, or cell adhesion (data not shown). A short list of eight genes was selected for further analysis due to their reported association with GC action and chondrocyte-mediated bone growth. Transcripts upregulated by Dex included lipocalin 2 (14.5-fold), CTGF (3.3-fold), integrin α10 (2.7-fold), SGCK (2.6-fold), and DMP1 (2.3-fold). Transcripts downregulated by Dex included Lumican (2.3-fold), secreted frizzled-related protein-2 (SFRP) (1.7-fold), and IGF-I (1.5-fold). To validate microarray data using independent methods, the eight Dex-responsive genes listed above were selected for analysis of gene expression by qRT-PCR. The induction of lipocalin 2, SGCK, CTGF, integrin α10, and DMP1 paralleled the expression patterns obtained by microarray analysis, although the fold change was much greater for qRT-PCR analysis (42-, 4.5-, 5.8-, 3.2-, and 5.7-fold, respectively; Fig. 1). Similarly, a 6.5-fold downregulation of IGF-I expression was detected by qRT-PCR analysis, although the downregulation of Lumican and SFRP seen in the microarray data was not confirmed. From these data, the expressions of six genes were confirmed to be altered by 24-h Dex treatment. We chose lipocalin 2 for further study because it had previously been shown to be upregulated by Dex in various cell types (25), indicating that this may be a generic mechanism by which GCs exert their antiproliferative effects. Furthermore, the role of lipocalin 2 in regulating chondrocyte function and bone growth is presently unknown. Lipocalin 2 expression was induced 42-fold by Dex in ATDC5 cells during both chondrogenesis and terminal differentiation (data not shown). Moreover, basal lipocalin 2 expression was similar during both phases of maturation in these cells. Consequently, ATDC5 cells in the chondrogenesis phase were used to study the effect of lipocalin 2 on GC-induced growth retardation.

Fig. 1.

Quantitative (q)PCR confirming changes in glucocorticoid-responsive chondrocyte genes in ATDC5 cells. Eight genes identified by Affymetrix microarray analysis were examined for changes in expression following 24-h dexamethasone (Dex) incubation (10−6 M) in ATDC5 cells at day 15 of maturation. See text for definitions. Data are expressed as means ± SE; n = 3. ***P < 0.001, **P < 0.01.

View this table:
Table 1.

Upregulated GC-responsive chondrocyte genes as revealed by microarray analysis

View this table:
Table 2.

Downregulated GC-responsive chondrocyte genes as revealed by microarray analysis

Characterization of Dex-induced lipocalin 2 expression in chondrocytes.

qRT-PCR analysis was used to examine the temporal effect of 10−6 M Dex on lipocalin 2 expression. No effect was seen at 1 h, whereas Dex significantly (P < 0.001) increased lipocalin 2 expression at all other time points studied (6–168 h), with an apparent plateau at 48 h (75-fold) (Fig. 2A). Consequently, 48-h Dex incubation was used for all future ATDC5 experiments. Dex-induced lipocalin 2 expression in ATDC5 cells was also found to be concentration dependent (Fig. 2B). At 10−6 M Dex, lipocalin 2 expression was 71-fold higher than control samples. This response was reduced to 44- and 13-fold at 10−7 and 10−8 Dex, respectively (P < 0.001), whereas no response was observed with 10−9 M Dex. Western blotting confirmed that 10−6 M Dex increased lipocalin 2 protein expression at 48 h in ATDC5 cells (Fig. 2D). Primary chondrocytes also expressed lipocalin 2, and this expression was increased with 10−6 M Dex treatment (3-fold at 6 h, P < 0.001). This increased expression was maintained for up to 72 h (Fig. 2C). Taken together, the results indicate that Dex regulates chondrocyte lipocalin 2 expression at both the mRNA and protein levels and that the regulation of gene expression is both time and concentration dependent.

Fig. 2.

Characterization of lipocalin 2 expression in chondrocytes. A: alteration in lipocalin 2 expression in chondrogenic ATDC5 cells treated with 10−6 M Dex for 1–168 h as measured by qPCR. Control, filled bars; Dex, open bars. B: dose-responsive change in lipocalin 2 expression in chondrogenic ATDC5 cells (day 8) with Dex over 48 h as measured by qPCR. C: expression of lipocalin 2 in primary murine chondrocytes treated with 10−6 M Dex for 6, 24, or 72 h as measured by qPCR. Control, filled bars; Dex, open bars. Data are expressed as means ± SE; n = 3. ***P < 0.001 vs. corresponding control. D: lipocalin 2 and β-actin protein expression in ATDC5 cells as measured by Western blotting after exposure to 10−6 M Dex for 48 h.

Lipocalin 2 expression in the presence of antiproliferative treatments.

To measure lipocalin 2 expression with antiproliferative treatments other than Dex (49), ATDC5 cells were cultured in the presence of ceramide (40 μM), TNFα (10 ng/ml), or in SFM for 48 h. Proliferation was significantly reduced in cells cultured in SFM, TNFα, and ceramide (88.8, 98.4, and 21.2%, respectively, P < 0.01; Fig. 3A). However, unlike Dex, lipocalin 2 protein expression was not induced in the presence of SFM or ceramide. A small increase in lipocalin 2 expression was observed with TNFα (Fig. 3B).

Fig. 3.

Lipocalin 2 expression with antiproliferative agents. A: proliferation in chondrogenic ATDC5 cells cultured in serum-free medium (SFM), or in the presence of TNFα (10 ng/ml) or ceramide (Cer.; 40 μM) for 48 h. Proliferation was measured by [3H]thymidine incorporation as previously described (n = 6; ***P < 0.001 vs. control, **P < 0.01 vs. control). B: lipocalin 2 and β-actin protein expression in ATDC5 cells as measured by Western blotting after exposure to SFM, TNFα (10 ng/ml) or ceramide (40 μM) for 48 h.

Immunolocalization of lipocalin 2 expression within the murine growth plate.

Immunohistochemistry revealed that, in metatarsals from 17-day-old fetal mice, lipocalin 2 was poorly expressed in the proliferating chondrocytes, although strong positive staining was evident in the prehypertrophic and hypertrophic zones (Fig. 4A). The pattern of lipocalin 2 distribution appeared to change in the postnatal growth plate. In the 10-day-old epiphysis, greater lipocalin 2 expression was localized to the proliferating chondrocytes (Fig. 4B), and this was more clearly seen in the 5-wk-old growth plate, where intense lipocalin 2 staining was observed in both the proliferating and hypertrophic zones (Fig. 4C). High levels of expression were also observed in the articular chondrocytes (data not shown) and within the chondrocytes of the developing secondary ossification center (Fig. 4D) of the 10-day-old tibias. All control sections incubated with normal rabbit serum were negative (Fig. 4E). The immunolocalization results clearly show that lipocalin 2 is expressed by murine growth plate chondrocytes and this confirms our qRT-PCR data from primary murine chondrocytes (Fig. 2C). In the epiphyses of older mice, lipocalin 2 appears to be present in all growth plate maturational zones. However, in the growth plates of younger mice, it is more restricted to the maturing hypertrophic cells.

Fig. 4.

Lipocalin 2 localization in murine proximal tibia and metatarsal. A: in metatarsals from 17-day-old fetal mice, lipocalin 2 is poorly expressed in the proliferating chondrocytes (P) and in the mineralizing zone (M); however, strong positive staining is present in prehypertrophic and hypertrophic chondrocytes (arrows) (H) (scale bar, 400 μm). B: in 10-day-old growth plate, lipocalin 2 expression is localized to proliferating (P) and prehypertrophic chondrocytes (PH). C: a similar localization was observed in 5-wk-old growth plate (scale bar, 200 μm), where intense lipocalin 2 staining is present in the proliferating (P), hypertrophic (H), and reserve zones (R) (arrows). D: lipocalin 2 is also present in hypertrophying chondrocytes of the developing secondary ossification center (arrows) of 10-day-old growth plate (scale bar, 200 μm). E: all control sections incubated with normal rabbit serum were negative.

Mechanism of Dex-induced lipocalin 2 expression in chondrocytes.

Incubating ATDC5 cells together with Dex and CHX caused a further increase in lipocalin 2 expression, suggesting that CHX was blocking the synthesis of a protein involved in lipocalin 2 repression (Fig. 5A). The competitive glucocorticoid receptor (GR) antagonist RU-486 caused lipocalin 2 expression to return to basal levels in the presence of Dex (Fig. 5B) (13-fold change in the presence of RU-486 and Dex vs. 39-fold change with Dex alone, P < 0.01). As with most steroidal competitive antagonists, RU-486 exhibits partial agonist properties under certain conditions (74), and this could explain the increase in lipocalin 2 expression with RU-486 alone (Fig. 5B).

Fig. 5.

Involvement of the glucocorticoid (GC) receptor (GR) in GC-induced lipocalin 2 expression in ATDC5 cells. A: lipocalin 2 expression with dexamethasone (Dex) plus the protein synthesis blocker cycloheximide (CHX). Chondrogenic ATDC5 cells (day 8) were preincubated for 30 min with CHX (8 μg/ml) before addition of 10−6 M Dex for 48 h. B: lipocalin 2 expression with Dex plus the GR antagonist RU-486. Chondrogenic ATDC5 cells (day 8) were incubated with 10−6 M Dex + RU-486 (10−5 M) for 48 h. Data are expressed as means ± SE; n = 3. ***P < 0.001 vs. control; αP < 0.05 vs. Dex.

Functional effects of lipocalin 2 on chondrocyte proliferation and differentiation.

To identify the possible physiological effect of Dex-induced lipocalin 2 expression on chondrocyte dynamics we determined, in initial studies, the effects of both lipocalin 2 overexpression (Fig. 6A) and recombinant lipocalin 2 on ATDC5 proliferation. Proliferation was significantly reduced (P < 0.001) in overexpressing cells (Fig. 6B) and those treated with recombinant lipocalin 2 (Fig. 6C). The reduction in proliferation by recombinant lipocalin 2 was modest (23.1% at 10 μg/ml), whereas a more dramatic effect (49.1%) was noted in the overexpressing cells. This difference is likely to be due to the unstable nature of synthetic recombinant proteins; therefore overexpressing cells were used for further analysis. Lipocalin-2 overexpression caused an increase in collagen type X expression (4-fold, P < 0.001; Fig. 7A). The effects of lipocalin-2 overexpression on chondrocyte proliferation (64% reduction vs. control cells, P < 0.001) and collagen type X expression (8-fold increase vs. control, P < 0.001) were further amplified by the addition of 10−6 M Dex (Figs. 6B and 7A). This apparent synergistic effect of Dex and lipocalin 2 on proliferation and collagen type X expression may be in response to the further increase in lipocalin 2 expression (140-fold, P < 0.001) noted in the Dex-treated overexpressing cells (Fig. 7B).

Fig. 6.

Effect of increased lipocalin 2 expression on chondrocyte proliferation. A: lipocalin 2 overexpression in ATDC5 cells was confirmed by PCR of lipocalin 2, puromycin resistance gene, and GAPDH (I); and Western blotting for lipocalin 2 and β-actin loading control (II). OE, lipocalin overexpressing; C, control cells containing empty vector only. B: proliferation of lipocalin 2-overexpressing cells as measured by [3H]thymidine incorporation. Transfected chondrogenic ATDC5 cells (day 8 of maturation) were incubated with or without 10−6 M Dex for 48 h. C: ATDC5 proliferation as measured by [3H]thymidine incorporation following 48-h incubation with recombinant lipocalin 2 protein (5–20 μg/ml, day 8 of ATDC5 maturation; n = 6). ***P < 0.001 vs. control.

Fig. 7.

Combined effect of lipocalin 2 and Dex on ATDC5 cell differentiation. A: type X collagen expression in lipocalin 2-overexpressing chondrogenic ATDC5 cells incubated with or without 10−6 M Dex for 48 h. B: effect of Dex on lipocalin 2 expression in chondrogenic ATDC5 cells overexpressing lipocalin 2. Cells were incubated with 10−6 M Dex for 48 h, and lipocalin 2 expression was measured by qPCR Data are expressed as means ± SE; n = 3–6. ***P < 0.001 vs. control; αP < 0.05 vs. Dex Empty; βP < 0.05 vs. lipocalin Control.


GCs are the most effective anti-inflammatory agents known. However, in children their long-term use leads to growth retardation through a combination of systemic effects through the GH-IGF axis and direct effects on the growth plate (3, 49). The exact mechanisms by which GCs exert their effects at the growth plate are still unclear but may be related to a reduction in chondrocyte proliferation combined with an increase in apoptosis, in addition to alterations in IGF-I signaling mechanisms (50). Murine in vivo studies carried out by us (unpublished observations), and others (60), have demonstrated that, after 1 wk of Dex treatment, the total width of the growth plate in 3-wk-old mice is significantly decreased due to a decrease in the width of the proliferative zone. In addition, we (10, 52) have previously used the murine chondrogenic ATDC5 cell line to show that, in vitro, Dex causes a significant reduction chondrocyte proliferation and an increase in alkaline phosphatase (ALP) activity, a known marker of terminally differentiating chondrocytes. In this study, we identified a number of chondrocyte genes up- or downregulated by Dex. Some of these, such as lipocalin 2, IGF-I, DMP1, and CTGF, have previously been shown to play an important role in the growth plate and were therefore selected for further analysis (29, 30, 63, 72).

We confirmed that CTGF, integrin α10, DMP1, SGCK, and lipocalin 2 were upregulated by Dex in ATDC5 cells. The Wnt antagonist SFRP-2, the proteoglycan Lumican, and IGF-I were all downregulated by Dex in the microarray, although only IGF-I was confirmed as being downregulated after qPCR analysis. Variations in results between microarray and qPCR analysis have been widely reported and is caused by a number of factors, including the microarray normalization technique used, measurement of spot intensity in the microarray, and tissue preparation between experiments (47). CTGF, a downstream mediator of transforming growth factor-β, is a critical growth factor for chondrocyte proliferation and differentiation, which has previously been shown to be upregulated by Dex in the chondrocytic HCS-2.8 cell line (38). DMP1-null mice display irregular and highly expanded growth plates, due to both a widening of the proliferation and hypertrophic zones, leading to a phenotype resembling dwarfism and chondrodysplasia (20). Interestingly, a recent study describing Affymetrix microarray on Dex-treated primary chondrocytes also found that DMP1 expression was significantly changed (31). The finding that Dex induces integrin α10 expression has not been previously reported. However, interestingly, it has recently been shown that integrin α10-null mice display growth retardation due to increased apoptosis of growth plate chondrocytes (4). SGCK is a serine kinase that has a catalytic domain homologous to that of Akt and can be activated by PI 3-kinase, making it an important signaling molecule in growth factor and insulin-dependent signaling pathways. SGCK has previously been shown to be upregulated in osteoblasts in response to Dex (40).

Lipocalin 2, also known as neutrophil-associated gelatinase lipocalin (NGAL), was originally identified as a 25-kDa protein that binds to small lipophilic substances such as bacteria-derived lipopolysaccharide (LPS). Since then, the known actions of lipocalin 2 have grown, and lipocalin 2 has now been linked with a wide range of different functions, including partuition, inflammation, differentiation, apoptosis, and acute-phase immunity. Lipocalin 2-null mice have compromised immunity against bacterial infection (6) but have no obvious skeletal phenotype (22). Originally termed SIP24 (superinducible protein-24), lipocalin 2 is known to be induced by a number of factors and conditions, such as LPS, CHX, serum, FGF2, prostaglandin F, oxidative stress, partuition, and Dex (43).

In this study, we have shown that lipocalin 2 expression is significantly increased with Dex treatment in murine chondrogenic ATDC5 cells. This increased expression was Dex concentration dependent and reached a plateau after 48-h Dex. Other antiproliferative treatments such as SFM and ceramide (46) did not cause an increase in lipocalin 2 expression, suggesting that the increase in lipocalin 2 expression by Dex observed in this study is not directly due to a reduction in proliferation. Previous studies have shown that lipocalin 2 expression is increased in the presence of proinflammatory mediators and inflammatory cytokines such as LPS (62, 73), TNFα, and IL-17 (58), which is in agreement with the small increase in lipocalin 2 expression observed with TNFα in this present study.

In primary murine costochondral chondrocytes isolated from the rib cages of 1-day-old mice, an increase in lipocalin 2 expression was also observed, and, although significant, this increase was smaller than that seen in ATDC5 cells. ATDC5 cells were originally isolated from a differentiating culture of AT805 teratocarcinoma cells, which display a fibroblast-like phenotype until the addition of insulin, which promotes their differentiation into chondrocytes (1). These cells are a useful model for studying chondrogenesis due to the fact that their temporal sequence of maturation directly mirrors the maturational phases within the growth plate. In addition, the cell population differentiates collectively, so that all the cells are at the same stage of differentiation at the same time. However, because of this, they cannot be compared directly to primary costochondral chondrocytes, which are isolated as a population of chondrocytes at all stages of differentiation. The fact that ATDC5 cells and primary costochondral chondrocytes comprise two different phenotypes could explain the differences in the magnitude of lipocalin 2 expression observed in this study. In addition, it has previously been shown that lipocalin 2 protein expression is increased during the neoplastic transformation of chondrogenic lineage cells (73), which may also contribute to the large increase in lipocalin 2 expression observed in ATDC5 cells.

The expression of lipocalin 2 in primary chondrocytes was supported by the finding that lipocalin 2 was expressed in the metatarsals of embryonic mice. Lipocalin 2 has previously shown to be expressed in the liver, spleen, lung, muscle, heart, and tibia of embryonic mice (25, 63), and within the tibia of embryonic rats, the rat homologue of lipocalin 2 neu-related lipocalin (NRL) is localized to the prehypertophic chondrocytes (73), supporting the findings of this study. It has previously been found that ALP has a similar pattern of localization within the growth plate (50), suggesting that, like ALP, lipocalin 2 may be a marker for chondrocyte differentiation. Interestingly, we have shown that, in 10-day-old and 5-wk-old mice, lipocalin 2 is more prominent in the proliferative chondrocytes, suggesting that the localization of lipocalin 2 expression within the growth plate is age dependent. Differences in lipocalin 2 expression between adult and fetal tissue have previously been described, with a general decline in expression in conjunction with aging (25). This, together with similarities in expression to ALP, and previous studies showing the pro-differentiating effects of lipocalin 2 in epithelial cells (71), further supports the hypothesis that lipocalin 2 may have a role in the regulation of chondrocyte differentiation.

In an attempt to identify some of the mechanisms involved in Dex-induced lipocalin 2 expression, the GR antagonist RU-486 was incubated alone or with Dex in ATDC5 cells. Results showed that, in the presence of RU-486 alone, lipocalin 2 expression was increased to some extent. RU-486 is a competitive antagonist for the GR, and, by nature, such antagonists often display some agonist activity at the receptor they competitively bind to, accounting for the increases in lipocalin 2 expression observed here. When RU-486 and Dex were incubated together, lipocalin 2 expression could no longer be induced with Dex. Analysis of the lipocalin 2 sequence with MatInspector (www.genomatix.de) confirmed previous reports of two GC response elements (GRE) in the 5′-regulatory region of the lipocalin 2 promoter (8, 25). The presence of GRE in the promoter region of lipocalin 2 suggested that this gene could be a direct target of the GR complex. This was supported by the fact that the Dex-induced increase in lipocalin 2 mRNA levels was not inhibited by CHX, a protein synthesis inhibitor. Interestingly, CHX in fact caused a further increase in lipocalin 2 expression. This finding has been previously reported (13) and could be due to CHX blocking the synthesis of a lipocalin 2 or Dex repressor.

The importance of lipocalin 2 in GC-induced growth retardation was further supported with the finding that its overexpression caused a reduction in ATDC5 proliferation and an increase in differentiation, an effect that is similar to that seen with GC-treated ATDC5 cells (49, 52). The effects on proliferation were reproduced with the addition of lipocalin 2 recombinant protein, again supporting the hypothesis that lipocalin 2 is influencing chondrocyte dynamics in a similar way to GCs. Interestingly, the effects of lipocalin 2 overexpression on proliferation and differentiation were exacerbated with Dex, and lipocalin 2 expression in transfected cells was increased more than would be expected from the additive effects of Dex and lipocalin 2 expression alone. The synergistic effect of Dex and LPS on lipocalin 2 expression has previously been described (65); however, these findings suggest that lipocalin 2 itself may play a role in mediating the effects of Dex. It is therefore possible that lipocalin 2 could be exacerbating Dex-induced growth retardation by interacting directly with Dex to inhibit chondrocyte proliferation and increase differentiation.

Lipocalin 2 has previously been shown to be regulated by Dex in dendritic cells (65) and fibroblast-derived murine L-cells (25); however, its stimulation in chondrocytes has not previously been reported. This study has shown that the acute-phase binding protein lipocalin 2 is markedly increased in chondrocytes in response to Dex and may play a role in mediating GC effects at the growth plate. It is widely acknowledged that GCs are likely to affect a wide range of complex regulatory mechanisms, including Akt and MAPK signaling (10), cell cycle signaling (9), and the NF-κB pathway (2). Although this paper has suggested that lipocalin 2 may be one mechanism involved in GC-induced growth retardation, it is likely that a number of other regulatory mechanisms are also involved. However, in this study we have identified that lipocalin 2 expression is altered in Dex-treated chondrocytes and that overexpression of lipocalin 2 results in a similar phenotype as that seen in Dex-treated cells. This suggests that the increased expression of lipocalin 2 by Dex has the potential to modify chondrocyte activities.


This work was funded by the Biotechnology and Biological Sciences Research Council.


We thank Richard Talbot (Roslin Institute) for critical analysis of the microarray data presentation and Elaine Seawright (Roslin Institute) for technical assistance. Microarray hybridizations were carried out by the MRC Human Genome Mapping Project Resource Centre, Cambridge, UK.


  • 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