The matrix metalloproteinases (MMPs) are a large multigene family of over 24 secreted and cell surface enzymes that process or degrade numerous extracellular or pericellular substrates. Their targets include nearly all extracellular matrix (ECM) components, other proteinases, proteinase inhibitors, clotting factors, growth factors, cell surface receptors, and cell–cell adhesion molecules (Yong et al., 2001). MMPs play a role in many cellular activities, including morphogenesis, survival, angiogenesis, inflammation, wound healing, and signaling (Sternlicht and Werb, 2001). Much of the information known about their biology has come from studies in vitro. A major challenge in understanding the biology of MMPs is the identification of their function in vivo. To date, little is known concerning the distribution and/or role of MMPs in developing organisms. Knockouts of MMPs and their natural inhibitors, the TIMPs, in mice have suggested that some MMPs have important developmental roles, for instance, MMP-9 and MMP-14 in skeletal development (Vu et al., 1998; Holmbeck et al., 1999). However, others show little or no phenotype possibly due to functional redundancy between family members. Studies in other organisms such as chick, zebrafish, Xenopus, and Drosophila have suggested that MMPs play an important role in tissue morphogenesis during development (Cai and Brauer, 2002; Jung et al., 2002; Zhang et al., 2003c; Page-McCaw et al., 2003). In fact, the first collagenase activity identified was from studies on tadpole tail resorption during metamorphosis (Gross and Lapiere, 1962). In Xenopus several MMPs have been cloned as seen in Figure 2 (Patterton et al., 1995; Stolow et al., 1996; Yang et al., 1997; Carinato et al., 2000; Jung et al., 2002). Most of the work looking at their function has concentrated on their role in tadpole tail resorption and other aspects of metamorphosis (Berry et al., 1998; Damjanovski et al., 2000, 2001; Jung et al., 2002). Of the Xenopus MMPs cloned to date, only ST3 (MMP-11), Col3 (MMP-13), and MMP-9 have been shown to be expressed early in development (Carinato et al., 2000; Damjanovski et al., 2000). Damjanovski et al. (2001) have carried out overexpression experiments using Xenopus transgenics in which they expressed ST3, Col4 (MMP-18) and mouse MT-MMP5 (MMP-24) during Xenopus development using the CMV promoter. All three led to lethality during late embryonic development. Other than these experiments, very little is known about the function of Xenopus MMPs during early development. We have begun to look at the role MMPs play during Xenopus embryogenesis. We present here the expression patterns of three new Xenopus MMP genes and show them to be expressed at different times during development and in restricted patterns in the neural crest, ventral ectoderm, and early myeloid/macrophage cells.
Cloning of New Xenopus MMP Genes
To investigate the role of MMPs in development, several new Xenopus MMPs were identified in the NCBI expressed sequence tag (EST) database and sequenced. The identification and naming of the clones as XMMP-14, XMMP-15, and XMMP-7 was based on their sequence homology with known MMPs in other species (Fig. 1), and by their position in a phylogenetic tree (Fig. 2). Like human MMP-7 and MMP-26, Xenopus MMP-7 lacks a hemopexin domain.
Temporal Expression Analysis of the Xenopus MMPs
The temporal expression pattern of these three metalloproteinases during Xenopus embryogenesis was first analyzed by reverse transcriptase polymerase chain reaction (RT-PCR) (Fig. 3). Expression of XMMP-14 was first detected in stage 10 embryos and stayed on until tadpole stages. XMMP-15 was present in all stages tested both as maternal and zygotic message. XMMP-7 expression was first seen at stage 17 and then increased through subsequent stages.
Expression of XMMP-14 (XMT1-MMP)
To further examine the role of these genes in development, their spatial expression pattern was visualized by whole-mount in situ hybridization. The earliest significant expression observed for XMMP-14 was from stages 17/18 in two stripes in the hindbrain, corresponding to two rhombomeres (Fig. 4A). These were identified as rhombomeres 3 and 5 by double-label RNA in situ analysis using the XMMP-14 probe and Krox-20, a marker for rhombomeres 3 and 5 and for neural crest (Fig. 4J). By stage 20/21, staining was still apparent in rhombomeres and in neural crest migrating from the rhombomeres (Fig. 4B). Expression was seen along the neural tube by stage 23 and in the trunk neural crest (Fig. 4C). By stage 26/27 (Fig. 4D,E), the staining was confined to a subset of the neural crest migrating in the branchial arches and also to ventral regions of the embryo. It also remains on in the rhombomeres. By stage 31, neural and branchial arch staining was no longer detectable, with some expression left in the tail bud and the proctodeum, which persisted until stage 36 (Fig. 4F,G). Sections of stage 18 and 23 embryos clearly show expression in the neural crest (Fig. 4H,I).
Expression of XMMP-15 (XMT2-MMP)
XMMP-15 showed low-level staining of the epidermis from stage 10 in the animal cap ectoderm (data not shown and Fig. 5G). By stage 17, this staining effectively covers the whole ectodermal surface of the embryo (Fig. 5A) but in later stages is confined to ventral regions (Fig. 5B–D). There was no staining of the non–animal cap tissue, which is destined to form mesoderm and endoderm. This finding was confirmed by sectioning (Fig. 5G,H). Expression of XMMP-15 was localized in small intense patches in ventral ectoderm (Fig. 5E, which is a magnified portion of Fig. 5D). This was also discernible at other stages (Fig. 5B, staining of the embryo shown was allowed to develop for longer, and it is consequently darker than in the other embryos). Although the ventral epidermis consistently expressed XMMP-15, it is notable that it clearly was absent from the fin and the tail (Fig. 5D,F,H).
Expression of XMMP-7
From sequence alignments, we conclude that EST 732291 is XMMP-7, the homolog of mammalian MMP-7 (matrilysin-1, see Fig. 1). The full sequence of XMMP-7 contained a pro sequence, catalytic domain, and, like mammalian MMP-7, lacked a hemopexin domain. XMMP-7 expression showed a remarkable speckling pattern due to expression in single cells. These were first seen in an anterior ventral area at stage 18; subsequently, the cells spread to cover the body by stage 30 (Fig. 6A–G). At stage 18, expression was visible in a small ventral patch of cells toward the anterior of the embryo in a region where the ventral blood islands form later (Fig. 6A). These cells spread along the anterior–posterior axis (Fig. 6B), before dispersing laterally and resolving into discrete spots by stage 22 (Fig. 6C). From there, the expression in the single cells continued to spread to cover the whole embryo in a spotted pattern by stage 36 (Fig. 6C–F), although a predominance for ventral expression continues up to stage 37/38 (data not shown). Sections showed that the expression was limited to the region between the epidermis and mesoderm (Fig. 6H–J). This staining pattern is reminiscent of that found for the peroxidase XPOX2 and a Ly6/uPAR-related protein (XLURP-1), which are expressed in Xenopus leukocytes (Smith et al., 2002). We, therefore, carried out double-label RNA in situ analysis with these probes to determine whether they were expressed in the same cells. Figure 6Gi shows XMMP-7 staining in blue and XPOX-2 in red. When the blue strongly overlaps the red color, the latter is hard to distinguish. However, Fast Red also fluoresces (Fig. 6Gii). Colocalization of the XMMP-7 and XPOX2 can then be clearly seen (arrowheads). The expression pattern of XMMP-7 is also similar to the expression of XMMP-9 (Carinato et al., 2000), although in that case, the cell type was not identified. Both mammalian MMP-9 and MMP-7 are known to be expressed in monocytes/macrophages (Busiek et al., 1992; Masure et al., 1993; Worley et al., 2003). The results of sequence alignments and expression, therefore, would support our conclusion that this EST clone is the Xenopus homologue of MMP-7, which is expressed specifically in tissue resident macrophages.
The data reported in this study pave the way for important studies on the roles of MMPs in two major developmental pathways, namely, neural crest formation and migration and monocyte/macrophage development and migration. Whereas certain MMPs previously have been implicated in chick and Xenopus neural crest migration (Alfandari et al., 1997, 2001; Cai et al., 2000), including MMP-2 (Cai et al., 2000; Duong and Erickson, 2004), the discovery of the expression of XMMP-14 (or MT1-MMP as it is also often named) in developing and migrating cranial and trunk neural crest is of great significance. This proteinase is directly responsible for the degradation of several ECM substrates, including fibronectin (important in neural crest migration) laminin and even native type I collagen (which is only degraded by a very limited number of collagenolytic enzymes). In addition, MMP-14 has the ability to initiate the activation process, whereby pro–MMP-2 is converted to its active form (Murphy et al., 1999); thus, that both genes are expressed in migrating neural crest may reflect this interaction. Future studies looking at the expression and function of XMMP-2 during neural crest formation will lead to a better understanding of this interaction and its importance in neural crest function. Both ST3 and Col3 are expressed in the branchial arches (Carinato et al., 2000; Damjanovski et al., 2000). It will be interesting to see how this expression is related to neural crest migration into these regions. Recently Sarras and colleagues have begun to look at the function of MMP-2, TIMP2, and MT-MMP in zebrafish development. Loss of function studies lead to embryos with severe axial deformation and an absence of well-developed cranial tissues (Zhang et al., 2003a–c). In mouse, the knockout of MMP-14 (MT1-MMP) results in death at approximately 1–3 months after birth, probably due to inadequate skeletal development (Holmbeck et al., 1999). No defect was reported to date in neural crest or its derivatives, but it is possible that this area was overlooked or that compensation may occur through expression of other MMPs. However, XMMP-14 is an extremely important enzyme, and functional studies in Xenopus will allow us to answer questions of redundancy much more readily and rapidly than in mouse.
As yet little is known about the function of MMP-15 (MT2-MMP). It is known to be expressed in epithelial cells in the endometrium, liver, and in carcinoma cell lines (Sato and Seiki, 1996; Theret et al., 1998; Zhang et al., 2000). The expression shown in Figure 2 suggests a possible function in the epidermis of the embryo.
MMP-7 (matrilysin-1) is a 28-kDa enzyme possessing catalytic activities against a broad range of ECM substrates, including proteoglycans, gelatin, fibronectin, laminin, and elastin (Busiek et al., 1992). It also has other substrates, including β4 integrin (von Bredow et al., 1997), syndecan-1 (Li et al., 2002), and procryptidins, which generate α-defensins (Shirafuji et al., 2003). MMP-7 has been shown to be expressed in human monocytes and macrophages and the epithelium of several noninjured, noninflamed tissues such as the lung, liver, and breast (Wilson et al., 1995). Due to a lack of mature α-defensins, MMP-7 null mice have an impaired ability to combat infections from enteric pathogens (Wilson et al., 1999). Transepithelial migration of neutrophils is impaired in MMP-7 null mice due to the absence of syndecan-1 cleavage (Li et al., 2002). Up-regulation of MMP-7 has been implicated in poor prognosis in human adenocarcinomas (Nakamura et al., 2002) and mammary tumour formation (Vargo-Gogola et al., 2002). In addition to being expressed in intact tissues, MMP-7 is also expressed in migrating epithelia. In fact, MMP-7 null mice have the most severe wound repair defect among MMP knockout mice generated to date (Wilson et al., 1999; Parks and Shapiro, 2001). The highly restricted expression of XMMP-7 in Xenopus embryonic macrophages as shown in this study suggests a role in their migration, from the ventral blood island where they form, throughout the whole embryo. The β4 integrins are mainly found in the basal surface of skin epidermal cells, and syndecan-1 is expressed in the Xenopus embryonic epidermis (Teel and Yost, 1996). Therefore, both are to be found in the area that the XMMP-7–expressing macrophages are to be found and so may be playing a role in their migration. XMMP-9 and ADAM-23 are also known to be expressed in these cells (Carinato et al., 2000; Smith et al., 2002), and it will be interesting to see what role they might have and if they could be interacting with XMMP-7.
In conclusion, targeted screening of the Xenopus EST database and RNA in situ hybridization has identified three MMPs with possible roles in Xenopus neural crest, epidermal and macrophage development respectively. Further investigation will begin to elucidate their functions in the development of these cells and tissues.
Reverse transcriptase-polymerase chain reaction (RT-PCR) analysis was performed using total RNAs from embryos at the indicated stages. The primers and PCR conditions were as follows: XMMP-7 forward 5′-gcc aat gcc cca acc tga ag-3′, reverse 5′-aat ctc cat gcg tgc gtg ctc-3′, annealing temperature 63°C, 30 cycles; XMMP-14 forward 5′-agc cgg aag agg atg atg aca-3′, reverse 5′-cca aca ggc cct aat aac caa t-3′, annealing temperature 53°C, 30 cycles; XMMP-15 forward 5′-ggg gtg cca aga aga agt ttt t-3′, reverse 5′-ctc ccc cat gtt agc acc ac-3′, annealing temperature 53°C, 30 cycles; Histone H4 forward 5′-cgg gat aac att cag ggt atc act-3′, reverse 5′-atc cat ggc ggt aac tgt ctt cct-3′, annealing temperature 55°C, 35 cycles.
Whole-mount in situ hybridization was carried out as previously described (Harland, 1991). Sense controls were carried out for most stages. Double-label RNA in situ hybridization was done according to (Knecht et al., 1995). Wax sectioning was carried out after staining to examine internal expression (Ciau-Uitz et al., 2000). Embryos were staged according to Nieuwkoop and Faber (1994).
We thank Stuart Smith and Tim Mohun for the XPOX-2 probe and Nancy Papalopulu for the Krox-20 probe and the efficiency of the University of Dundee Sequencing service. We also thank Andrea Münsterberg and Dylan Edwards for their comments on the manuscript. T.G. received an MRC studentship, and C.Y. received a Company of Biologists Summer Studentship.