Glucocorticoids Alter Craniofacial Development and Increase Expression and Activity of  Matrix Metalloproteinases in Developing Zebrafish (Danio rerio)
Jedd M. Hillegass, Caren M. Villano, Keith R. Cooper and Lori A. White1Author AffiliationsJoint Graduate Program in Toxicology, Rutgers, The State University of New Jersey, NewBrunswick, New Jersey 08901
Teratogenic effects are observed following long-term administration of glucocorticoids,although short-term glucocorticoid therapy is still utilized to reduce fetal mortality, respiratorydistress syndrome, and intraventricular hemorrhage in preterm infants. However, the mechanism of glucocorticoid-induced teratogenicity is unknown. We hypothesize that glucocorticoid-inducedteratogenesis is mediated through the glucocorticoid receptor (GR) and results from altering theexpression and activity of the matrix metalloproteinases (MMPs). During embryogenesis,degradation of the extracellular matrix to allow for proper cellular migration and tissueorganization is a tightly regulated process requiring appropriate temporal and spatial expressionand activity of the MMPs. Studies have demonstrated that MMP gene expression can be either inhibited or induced by glucocorticoids in a variety of model systems. Using the zebrafish (Daniorerio) as a model of development, the data presented here demonstrate that embryonic exposure tothe glucocorticoids dexamethasone or hydrocortisone increased expression of two gelatinases,MMP-2 (
1.5-fold) and MMP-9 (7.6- to 9.0-fold), at 72 h postfertilization (hpf). Further,gelatinase activity was increased approximately threefold at 72 hpf following glucocorticoidtreatment, and changes in craniofacial morphogenesis were also observed. Cotreatment of zebrafish embryos with each glucocorticoid and the GR antagonist RU486 resulted in attenuationof glucocorticoid-induced increases in MMP expression (52–84% decrease) and activity (41–94%decrease). Furthermore, the abnormal craniofacial phenotype observed following glucocorticoidexposure was less severe following RU486 cotreatment. These studies demonstrate that in theembryonic zebrafish, dexamethasone, and hydrocortisone alter expression and activity of MMP-2and -9, and suggest that these increases may be mediated through the GR.Glucocorticoids play a central role in vertebrate physiology and are involved in numerousregulatory mechanisms associated with development, bone replication, bone differentiation,apoptosis, metabolism, circadian cell cycle rhythmicity, and the stress response, among others(Canalis and Delany, 2002; Dickmeis et al., 2007). Glucocorticoids act as immunosuppressive andanti-inflammatory agents, and are therefore widely utilized to treat autoimmune and inflammatorydisorders, transplant rejection, and lymphoproliferative diseases (Almawi et al., 2002). Thesecompounds are also potent teratogens whose antenatal use has been linked to fetal growthrestriction and cleft palate (Abbott, 1995; Mandl et al., 2006), and intrauterine programming of metabolic, neuroendocrine and cardiovascular disorders in adult life (Seckl, 2004). The exact
mechanism by which glucocorticoids exert their teratogenic effects is unknown, although recentwork in the zebrafish has demonstrated that these compounds are capable of upregulating matrixmetalloproteinase-13 (MMP-13) (Hillegass et al., 2007). Further, glucocorticoids have been shownto stimulate MMP-9 expression through induction of soluble glucocorticoid-induced tumor necrosis factor receptor in murine macrophages (Lee et al., 2003, 2004). These results sit in directopposition to what has been found in mammalian cell models in vitro, where MMP expression isinhibited by exposure to glucocorticoids (Canalis and Delany, 2002; Chakraborti et al., 2003;Vincenti et al., 1996). Despite contradictory findings, these studies suggest MMPs may be pivotalin eliciting glucocorticoid-induced teratogenicity and support further examination of themechanism.Extracellular matrix (ECM) remodeling is essential for a number of physiological processes including embryonic development, reproduction, tissue resorption, wound healing, andapoptosis (Brinckerhoff and Matrisian, 2002; Hulboy et al., 1997; Nagase and Woessner, 1999).Central to these processes are MMPs, a group of over twenty zinc-dependent endopeptidasesresponsible for precise and regulated ECM degradation. Dysregulation and excessive expression of MMPs have been tied to a number of pathological disorders such as osteo- and rheumatoidarthritis, emphysema, multiple sclerosis, bacterial meningitis, and tumor invasion and metastasis(D'Armiento et al., 1992; Folgueras et al., 2004; Hendrix et al., 2003; Leppert et al., 2001;Rundhaug, 2005). Recent work using zebrafish has focused on the role of MMPs during embryonicdevelopment. Our laboratory and others have shown that MMP-2, MMP-9, MMP-13, andmembrane-type MMPand -β are required for normal zebrafish embryogenesis (Hillegass et al.,2007, unpublished data; Zhang et al., 2003a, b).Most glucocorticoid-associated effects are mediated through the glucocorticoid receptor (GR), which belongs to the nuclear receptor superfamily and acts as a ligand-dependenttranscription factor (Evans, 2005). A single GR has been identified in zebrafish thus far, althoughother related teleosts have two distinct GR genes (Bury et al., 2003; Greenwood et al., 2003). Theteleost GR is unique from the mammalian GR in that the DNA-binding domain contains nineadditional residues between the two zinc fingers. These nine amino acid inserts are remarkablyconserved among teleostean fish species (Stolte et al., 2006; Terova et al., 2005) and appear to bethe result of alternative splicing (Stolte et al., 2006). It has been suggested that because theseresidues promote greater DNA affinity in the GR, they could have been selected to serve the largespectrum of cortisol functions in fish (Lethimonier et al., 2002). Such GR splice variants have beencharacterized in the rainbow trout (Bury et al., 2003) and Burtons' mouthbrooder (Greenwood etal., 2003; Takeo et al., 1996) thus far, and it is believed this could result in separate biologicalfunctions for each receptor variant (Prunet et al., 2006).The purpose of these studies is to examine the effects of the glucocorticoidsdexamethasone and hydrocortisone on MMP-2 and MMP-9 expression and activity and toestablish a causal link between activation of the GR and changes in MMP-2, MMP-9, and MMP-13 levels. Further we wish to better characterize the craniofacial defects known to result from
exposure of embryonic zebrafish to these glucocorticoids. The gelatinases MMP-2 (gelatinase A)and MMP-9 (gelatinase B) readily degrade gelatins (denatured collagens) and intact collagen typeIV, and the collagenase MMP-13 (collagenase-3) cleaves native interstitial collagens I, II, and III(Chakraborti et al., 2003). These particular MMPs were selected because they have been shown to be vital for normal development in both zebrafish and mice (Inada et al., 2004; Itoh et al., 1997;Mosig et al., 2007; Stickens et al., 2004; Vu et al., 1998). The data presented in this paper demonstrate that dexamethasone and hydrocortisone cause increases in MMP-2 and MMP-9expression and activity in the zebrafish embryo at 72 h postfertilization (hpf), with resultantchanges in craniofacial morphogenesis. Cotreatment with glucocorticoids and the GR antagonistRU486 results in attenuation of the increases in MMP-2, MMP-9, and MMP-13 expression andactivity normally observed following glucocorticoid treatment, as well as a partial rescue of theabnormal craniofacial phenotype. These results further demonstrate that in the embryoniczebrafish, dexamethasone, and hydrocortisone alter expression and activity of MMPs, includingMMP-2 and -9, and suggest that these increases may be mediated through the GR.MATERIALS AND METHODSZebrafish strains and husbandry.The AB strain of zebrafish (Danio rerio), obtained from the Zebrafish InternationalResource Center, was used for all of the experiments described. Zebrafish were maintained and bred in an Aquatic Habitats recirculation system according to a husbandry protocol approved bythe Rutgers University Animal Care and Facilities Committee.
Collection and treatment of zebrafish embryos.
Embryos were collected from breeding stocks of zebrafish and treated starting at 3 hpf.Treatments consisted of continuous (i.e., without renewal) exposure through 24, 48, 72, or 96 hpf to 100 mg/l of dexamethasone (254.81μM equivalent) or hydrocortisone (275.88μM equivalent)alone or in conjunction with 100–250nM mifepristone (RU486) in embryo medium (Westerfield,2000). This dose of dexamethasone or hydrocortisone has been shown in our laboratory to beeffective in inducing MMP-13 levels in developing zebrafish (Hillegass et al., 2007).Dexamethasone and hydrocortisone were dissolved in dimethylformamide (DMF) prior to beingdiluted in embryo medium to the experimental concentrations. In order to generate experimentalconcentrations of RU486, a 1mM stock of RU486 was first made by dissolving in 100% ethanol,and a subsequent 100μM working stock was generated by further dilution into ethanol. Finalconcentrations of RU486 were achieved by diluting this working stock in embryo medium. AllRU486 stocks were stored at −20°C between uses. Controls, including a no treatment control of embryo medium alone and a solvent control consisting of ethanol and/or DMF alone were runconcurrently with each treatment. When cotreatments of dexamethasone or hydrocortisone andRU486 were performed, a RU486 control consisting of the appropriate concentration of RU486alone was included. The no treatment control embryos were monitored throughout the course of 
each study to confirm embryo viability. Data from any treatment in which > 10% of the solvent or no treatment control embryos exhibited developmental abnormalities were not considered for analysis. All chemicals and solvents were purchased from Sigma-Aldrich (St Louis, MO) and possessed purities ≥ 98%.
 Real-time reverse transcription–PCR.
Embryos collected for RNA isolation (n = 50–100) were snap frozen in 1.5-mlmicrocentrifuge tubes using liquid nitrogen and stored at −80°C. Total RNA was isolated fromembryos using TRIzol (Invitrogen Carlsbad, CA) and DNase-treated (DNA-free kit, AmbionAustin, TX) to remove genomic DNA contamination. Total RNA yields were typically 1–2 μg/μl.Reverse transcription was performed on 1 μg aliquots of total RNA to produce complimentaryDNA (cDNA) for real-time reverse transcription–PCR (RT-PCR) (quantitative RT-PCR; qRT-PCR) using an iScript cDNA Synthesis Kit (BioRad Hercules, CA). Real-time RT-PCR reactionswere performed in triplicate using BioRad iQ SYBR Green Supermix, and cDNA amplificationwas performed for 40 cycles on a BioRad iCycler equipped with an iCycler iQ Detection System.Primers to zebrafish β-actin, MMP-2, MMP-9, and MMP-13 were used in amplification reactions.For β-actin, the forward primer was 5′-CGAGCAGGAGATGGGAACC-3′ and the reverse primer was 5′-CAACGGAAACGCTCATTGC-3′ giving a product size of 102 base pairs (bp). For MMP-2, the forward primer was 5′-AGCTTTGACGATGACCGCAAATGG -3′ and the reverse primer was 5′-GCCAATGGCTTGTCTGTTGGTTCT-3′ giving a product size of 224 bp. For MMP-9, theforward primer was 5′-AACCACCGCAGACTATGACAAGGA-3′ and the reverse primer was 5′-GTGCTTCATTGCTGTTCCCGTCAA-3′ giving a product size of 89 bp. For MMP-13, theforward primer was 5′-ATGGTGCAAGGCTATCCCAAGAGT-3′ and the reverse primer was 5′-GCCTGTTGTTGGAGCCAAACTCAA-3′ giving a product size of 289 bp. Real-time thresholdcycle data were normalized to β-actin, which served as a loading control, and standard curvesgenerated for MMP-2, MMP-9, and MMP-13 were used to quantify messenger RNA (mRNA)expression.
 In situ hybridization.
The MMP-2 RNA probe was generated using a construct generously donated by Dr RobertTanguay (Oregon State University) consisting of full-length MMP-2 cDNA expressed in pCR-Blunt II-TOPO. The plasmid was linearized using PstI, and SP6 and T7 RNA polymerase wereused to generate antisense and sense, respectively, digoxigenin (DIG)-labeled RNA probes (DIGRNA Labeling Kit-SP6/T7, Roche Indianapolis, IN). The MMP-9 RNA probe was generated froma cDNA clone encoding a 318-bp portion of the MMP-9 gene amplified using the following primers: 5′-TTTGAGCTCTACAGTCTGTTTCTGGTGG-3′ (forward primer; the italicized portiondesignates a SacI restriction site) and 5′-ATAGGATCCGGCGTCAAACTCCTT-3(reverse primer; the italicized portion designates a BamHI restriction site). To generate this portion of theMMP-9 gene, total RNA from untreated 72 hpf embryos was isolated, DNase-treated, made intocDNA as described previously and polymerase-amplified via PCR. The 318-bp PCR product was
of 22