Journal of Cell Science 112, 3847-3853 (1999) 3847 Printed in Great Britain © The Company of Biologists Limited 1999 JCS0725 COMMENTARY The tenascin-C knockout revisited Eleanor J. Mackie1 and Richard P. Tucker2,* 1School of Veterinary Science, University of Melbourne, Parkville, Victoria, Australia 2Department of Cell Biology and Human Anatomy, University of California at Davis, Davis, California 95616, USA *Author for correspondence ([email protected]) Published on WWW 3 November 1999 SUMMARY In the past seven years, two groups have independently healing after suture injury of corneas. In both skin and produced tenascin-C-knockout mice. These mice are born corneal wounds, fibronectin expression is abnormally low alive and, originally, were described as showing no in tenascin-C-knockout mice. Finally, in vitro studies abnormalities. More recent studies, many involving indicate that haemopoietic activity is defective in bone pathological intervention, have shown that tenascin-C- marrow from these mice. When examined together, these knockout mice have several defects. The mice exhibit studies provide evidence for precise functions for tenascin- abnormal behaviour, as well as abnormalities in brain C, as well as an explanation for why the sequence of chemistry. They also show defects in structure and repair tenascin-C is so highly phylogenetically conserved. of neuromuscular junctions, in the ability to recover from snake-venom-induced glomerulonephritis and in chemically induced dermatitis. Healing of skin wounds is Key words: Homologous recombination, Phenotype, Extracellular morphologically normal, but the mice exhibit defects in matrix INTRODUCTION found almost exclusively in the central nervous system (CNS), and tenascin-X, a mammalian tenascin found in some In the early 1980s several labs independently discovered and connective tissue and around blood vessels. Tenascin-C was characterized the glycoprotein that eventually became known cloned from zebrafish, amphibians, birds and mammals, and its as tenascin. Immunohistochemistry revealed tenascin in highly conserved sequences provided further evidence for its everyone’s favorite extracellular matrix: around motile cells, fundamental function (Fig. 1; Chiquet-Ehrismann et al., 1994a; at hot spots of proliferation and at sites of branching Erickson, 1994). The precise function(s), however, remained morphogenesis, near inductive events, in cartilage, tendons elusive. Cell culture studies showed that tenascin-C promotes and quite prominently in the developing nervous system. both adhesion and detachment, and can stimulate and inhibit Tenascin was also shown to be upregulated at the margins of cell division, and that its ability to bind to classic integrins healing wounds and in the stroma of many tumours (for depends on the species from which the tenascin-C was isolated reviews see Erickson, 1993a; Tucker, 1994; Chiquet- (for reviews see Erickson, 1993a; Faissner et al., 1994). Some Ehrismann, 1995; Mackie, 1997). The results of in vitro of the differences are probably related to the type of cell under studies were just as titillating. In contrast to most substrate- investigation or the bioassay used. It also seemed likely that at adhesion molecules, purified tenascin has anti-adhesive least some of the contradictory observations would be properties. Most cells can attach weakly to tenascin-coated explained by studies of the many tenascin-C splice variants substrata, but they remain rounded and particularly motile (e.g. Mackie and Tucker, 1992). Many of us were shocked to (e.g. Lotz et al., 1989; Halfter et al., 1989). Tenascin was learn that tenascin-C-knockout mice, among the first to be touted as a ‘magic bullet’, potentially responsible for key made by homologous recombination, were indistinguishable events in systems as diverse as neural-crest morphogenesis from their wild-type littermates (Saga et al., 1992): the and tumour metastasis. tenascin-C-knockout mice were the same size as the controls; The cloning and sequencing of tenascin eventually revealed they were fertile; and cursory histological examinations a family of molecules in which tenascin’s distinctive EGF-like revealed no gross deficits in neuroarchitecture or principal repeats and fibrinogen-like C terminus are separated by a series organ systems (Fig. 2). Not only had gene-knockout of fibronectin type III repeats (Fig. 1; reviewed by Chiquet- technology failed to clarify a function for tenascin-C, it pointed Ehrismann et al., 1994a). The tenascin discovered first was to a minor or even completely redundant role for this protein named tenascin-C to distinguish it from tenascin-R, which is (Erickson, 1993b). 3848 E. J. Mackie and R. P. Tucker EGF-like repeats fibronectin type III repeats fbg-like domain A Fig. 2. An adult female tenascin-C-knockout mouse. Although phenotypically normal, tenascin-C-knockout mice have abnormal behaviour and responses to stress and trauma. B knockout mice. Forsberg et al. (1996) settled the controversy once and for all. They independently constructed a second Mouse KVEGYSGTAGDSMNYHNGRSFSTYDKDTDSAITNCALSYKGAFWYKNCHR tenascin-C-null mouse that was, like the first, phenotypically Human .............A.........F.....................R.... normal. Pig .............A.........F.......................... Chicken R.D..........T.........F...N...................... Evidence for the upregulation of other forms of tenascin to Xenopus ...R.............N...................... Zebrafish H.G..........T..H..P.....N.N.I.V.................. compensate for the missing tenascin-C has not been forthcoming. Saga et al. (1992) considered this possibility and looked for upregulation of tenascin-X (which they called Mouse VNLMGRYGDNNHSQGVNWFHWKGHEYSIQFAEMKLRPSNFRNLEGRRKRA Human .........................H........................ tenascin-MHC) by northern blotting: the levels of tenascin-X Pig ..........S....................................... mRNAs appeared to be as those in the wild-type mice. Cell- Chicken ......................................S........... Xenopus .........TS....I..................... culture studies (Sakai et al., 1996) and immunohistochemical C Zebrafish ..I.......S..K...........H.VE.....I..A....F...K..S analysis of tenascin-R in the brains of tenascin-C-knockout Fig. 1. (A) Murine tenascin-C is a modular glycoprotein, composed mice (Steindler et al., 1995) have yielded similar results. Even of 14 EGF-like repeats (red diamonds), a series of fibronectin type III some of tenascin-C’s most adamant supporters have concluded repeats (blue ovals) and a globular fibrinogen-like domain at its C from these studies that mice lacking tenascin-C do indeed terminus. One or more additional fibronectin type III repeats are develop in at least a grossly normal fashion and that, if found in some alternatively spliced isoforms (yellow oval). anything is compensating for the loss of tenascin-C function in (B) Tenascin-C exists in the extracellular matrix as a hexamer in these mice, it is probably not one of the other known tenascins. which six chains are covalently linked near their N-termini. We’re back to square one: what does tenascin-C do? Steps (C) Putative functional domains within tenascin-C are highly toward understanding non-redundant functions of tenascin-C, conserved between species, which implies fundamental roles in and explanations for its phylogenetic conservation, have finally development, disease or responses to stress. Here, the 100 residues at come from more-detailed analysis of the behaviour of the the C terminus of mouse (GenBank accession no. JQ1322), human (NP 002151), pig (S19694), chicken (212748) and zebrafish knockout mice and their responses to trauma (Table 1). (CAA61489) tenascin-C are aligned. The partial sequence of a Xenopus tenascin-C sequence (I51647) is included. The human and pig sequences share 99% similarity with mouse, the chicken 98%, TENASCIN-C-KNOCKOUT MICE HAVE ABNORMAL Xenopus 95%, and zebrafish 91%. BEHAVIOUR Tenascin-C-knockout mice are indeed born alive, fertile and THE FIRST STEP OF KNOCKOUT GRIEF: DENIAL outwardly normal. Their behaviour, however, is markedly abnormal. Fukamauchi et al. (1996) first reported this abnormal The knockout results were sufficiently unexpected that many behaviour four years after the development of the knockout. researchers suspected that the mice described by Saga et al. The delay was largely due to the time required to backcross the (1992) still express some form of tenascin-C or that the animals into a strain that was suitable for such analysis. Normal tenascin-R and tenascin-X are upregulated and compensate for C57BL/6N mice, as well as heterozygous littermates, can swim tenascin-C in the knockouts. Mitrovic and Schachner (1995) for three minutes in a water tank without training. In contrast, published results supporting the former hypothesis. Using many of the tenascin-C-knockouts fail the test and must be immunohistochemistry and in situ hybridization, they detected rescued before drowning. The mutant mice also have aberrant tenascin-C immunoreactivity and transcripts in the brains of behaviour on dry land. Normal mice move about their cages the tenascin-C-knockout mice. The signal was reduced in most frequently during the dark cycle. The tenascin-C- comparison with that evident in wild-type animals, and the knockout mice move about their cages almost incessantly, immunoreactivity appeared to be intracellular and not regardless of the dark-light cycle. In a follow up study extracellular, but the authors reasonably concluded that