The biology of thrombospondins and other modulatory extracellular matrix proteins
N. Anilkumar and Soren Prag
Page 1 - Page 2
The thrombospondin type 1 repeat (TSR) superfamily
Over the last couple of years, many new additional proteins that contain TSR have been
identified. The question of whether they have properties in common with
thrombospondin-1 and thrombospondin-2 is of more than academic interest
because of the anti-angiogenic activities of the TSR of thrombospondins.
A second theme at the meeting thus related to the biological roles of
these new relatives.
In her Plenary
Talk, Judith Kimble (University of Wisconsin, USA) gave an overview
of her labs work on GON-1 metalloprotease in C. elegans.
The gon-1 gene encodes a secreted metalloprotease with a metalloprotease
domain and multiple TSRs. The TSRs are postulated to anchor GON-1 to the
extracellular matrix and thereby localise its metalloprotease activity.
In gon-1 mutants, gonadogenesis is impaired such that no arm elongation
occurs and somatic gonadal structures are severely malformed. Kimble described
how Gon-1 is expressed in two sites, leader cells and muscle, and expression
in each site has a unique role in forming the gonad. The gon-1 gene is crucial for migration of the gonadal leader cells. In wild-type
animals, the leader cells migrate hundreds of microns, whereas in gon-1 null mutants, the leader cells do not move at all. The gon-1 mutants also
have defects in gonadogenesis that are unrelated to leader cell migration,
suggesting that gon-1 might have a second role in gonadogenesis.
this talk, it was interesting to hear the effects of a TSR relative being
knocked-out in mice. Kauji Kuno (Kanazawa, Japan) described the
work in his lab on a matrix-anchored metalloprotease, ADAMTS-1. He reported
that the precursor and processed forms of ADAMTS-1 are secreted from cells
and bind to the extracellular matrix (ECM) in a heparin-sensitive manner.
Analyses of deletion mutants of the ADAMTS-1 protein revealed that a spacer
region and the three TSR were important for a tight interaction with the
ECM. Targeted disruption of the mouse ADAMTS-1 gene resulted in growth
retardation with malformation of adipose tissue, kidneys and the adrenal
medullary gland. Interestingly, in
relation to the gon-1 mutants, female mice had malformed
ovaries and low fertility. ADAMTS-1 thus appears necessary for normal
growth, female gonadogenesis and the morphology and function of certain
Luisa Iruela-Arispe (UCLA, USA) described studies on the secretion and processing of ADAMTS1.
She showed that human ADAMTS-1 is secreted, then processed in two consecutive
steps to release two active forms. Matrix metalloproteinases 2, 8 and
15 were able to release the smaller form from the larger one. This second
processing step removes two TSR from the carboxyl terminal end of ADAMTS-1
and alters the affinity of the protein for heparin. ADAMTS-1 has anti-angiogenic
activity in culture assays, and she hypothesises that processing could
be relevant for modulating the anti-angiogenic properties.
Sunil Apte (Cleveland Clinic, USA) gave an overview of the whole ADAMTS family, of
which there are now around 25 members, and introduced a new subfamily
of TSR proteins, the ADAMTS-like (ADAM-TSL) proteins. This family includes
Punctin, a novel extracellular matrix component of muscle tissues.
There were two
interesting talks that related to the structure of TSR. The first NMR-structure
of a TSR-containing protein - heparin-binding growth-associated molecule
or HB-GAM was presented by Ilkka Kilpeläinen (Helsinki, Finland).
HB-GAM is an extracellular matrix-associated protein implicated in the
development and plasticity of neuronal connections in the brain which
contains two repeats distantly related to TSR at primary sequence level.
The NMR structure shows two domains of antiparallel b-sheets
separated by a long flexible hinge. This structure opens the way to new
studies of TSR and their functions.
There has been
some debate about whether TSR bind heparin directly. An interesting talk
from Jan Hofsteenge (FMI, Switzerland) cleared up some of the contradictions.
He described a new protein modification, C-mannosylation
of tryptophan residues, in which carbohydrate is attached to the indole
moiety of tryptophan within a specific recognition sequence. This modification
is not made in plants, yeast or eubacteria. Analysis of several TSR revealed
multiple C-mannosylation sites within the TSR. The functional significance
of this modification remains mysterious, yet when using synthetic peptides
in comparison to native protein, one should be aware of the possibilities
for different properties due to the absence of these post-translational
actions of modulatory matrix proteins
The cell biology
and signalling mechanisms initiated by thrombospondins emerged as another
hot topic of discussion in the meeting. The diversity in cellular responses
to thrombospondin-1 is explained by differential expression by different
cell types of multiple thrombospondin-1 receptors that induce distinct
signals. David Roberts (NIH, USA) reported that integrin a3b1 is
an activation-dependent adhesion receptor for thrombospondin-1 in breast
carcinoma cells and small cell lung carcinoma cells. This integrin is
also expressed in endothelial cells but its ability to recognize thrombospondin-1
is suppressed in confluent cells. When activated in subconfluent cells,
it contributes to cell spreading on TSP-1. Bill Frazier (Washington
University, USA) discussed the role of CD47 in cross-talk to integrin
a2b1. He described how activation of a2b1 by thrombospondin-1 or a CD47-binding
peptide stimulate smooth muscle migration by activation of a heterotrimeric
The co-organiser, Jo Adams (UCL, UK), described cellular responses to thrombospondin-1.
In contrast to fibronectin, collagen and various other extracellular matrix
proteins, cell adherence to thrombospondin-1 promotes the formation of
actin-containing microspikes. These microspikes contains the actin-bundling
protein, fascin and are negatively regulated by protein kinase C-dependent
phosphorylation of fascin. The formation of these spikes depends on activities
of Rac and Cdc42, but not of Rho GTPase. Mark Ginsberg (TSRI, USA)
presented a poster of recent data describing how the a4 integrin cytoplasmic
domain binds to paxillin. Fusion of the a4-tail to the aIIb-integrin resulted
in a reduction in the formation of focal adhesion and stress fiber formation,
indicating that the association of paxillin and a4 promotes integrin signalling
that leads to focal adhesion disassembly.
Paul Bornstein (University of Washington, USA) presented interesting studies of thrombospondin-2
null fibroblasts. These skin fibroblasts show altered adhesion properties
in culture. Compared to wild-type cells, the KO cells show increased activity
of metalloproteinase-2 (gelatinase-A). When KO cells are treated with
the specific gelatinase-A inhibitor, TIMP2, or with neutralising antibodies
against metalloproteinase 2, they regain full spreading on several matrix
Alan Rapraeger (Madison, USA) discussed the mechanisms by which heparan sulphate chains
on proteoglycans mediate stable high affinity binding of fibroblast growth
factors (FGFs) to their receptors and regulate FGF activity. Basement
membrane and cell surface heparan sulphate proteoglycans differ in their
ability to bind individual FGF family members, presumably due to domain
structure differences in the heparan sulfate chains.
SPARC and New Modulatory Proteins
other major theme of the meeting was advances in understanding the biology
and mechanisms of other modulatory extracellular matrix proteins. Ruth
Chiquet-Ehrismann (FMI, Switzerland) reported a novel family of proteins,
teneurins, which are predominantly expressed in the nervous system. Chicken
teneurin-1 and teneurin-2 were found as two homologues of the Drosophila pair-rule gene product Ten-m and Drosophila Ten-a. By in situ hybridisation, they showed that teneurin-1 and -2 are expressed by distinct
populations of neurons during the time of axonal growth, most prominently
in the developing visual system. The expression of teneurin-2 in neuronal
cells leads to a significant increase in the number of filopodia and to
the formation of enlarged growth cones. These proteins might have a function
in the formation of neuronal connections.
Jean Schwarzbauer (Princeton, USA) discussed cellular effects of tenascin-C in a matrix
gel designed to provide a matrix similar to that of healing wounds. Inclusion
of tenascin-C in the gel alters actin cytoskeletal organisation in cells
by modulating Rho GTPase activity.
Helen Sage (Washington, USA) described the phenotype of SPARC-null mice. A prominent
phenotype is aberrant connective tissue responses to injury. In culture,
cells from these mice exhibit significantly enhanced rates of proliferation.
Most strikingly, all strains of mice lacking the sparc gene develop early-onset
of cataracts associated with aberrant proliferation and differentiation
of lens epithelial cells. Additional phenotypic alterations associated
with aging, connective tissue, wound healing, and/or angiogenesis suggest
unanticipated roles for SPARC in the maintenance of certain organs and
tissues through life.
In his stimulating
talk, Matthias Chiquet (Bern, Switzerland) presented data about
the regulation of tenascin-C and collagen XII production by mechanical
stress. It is well known that connective tissue cells adapt their ECM
to changes in mechanical load. A feedback mechanism must exist, by which
cells that sense mechanical stress from their matrix respond by an altered
pattern of protein expression, and thus remodel the ECM to adapt to changing
mechanical requirements. The production of tenascin-C and collagen XII,
ECM proteins typical of tendons and ligaments, was found to be high in
fibroblasts attached to a stretched collagen matrix, but suppressed in
cells on a relaxed matrix. The response to a change in stress was rapid
and reversible, and was reflected at the mRNA level. Both the tenascin-C
and the collagen XII gene promoters contain stress-responsive
enhancer regions with similarity to shear stress response elements
found in other genes.
Richard Tucker (Davis, USA) gave us an entertaining and educational talk about his work
on tenascin-Y, the avian homologue to the mammalian tenascin-X. Tenascin-Y
is localised with the ependymal cells in the spinal cord and in the brain.
In the adult, the protein levels are high at boundaries between the Schwann
cells in the peripheral nervous system and the central nervous system.
Tenascin-Y inhibits Schwann cell motility in vitro and thus might define
a boundary in vivo.
The meeting ended
with dinner and dancing to a local jazz band. This second ad-hoc meeting
on modulatory matrix proteins generated a lot of enthusiastic discussion
amongst the participants and we anticipate will be continued in a series
for Molecular Cell Biology and Dept. of Biochemistry and Molecular
University College London
Gower Street, London WC1E 6BT, UK.
1 - Page 2
BSCB Hooke Medal Winner
Cool pics wanted
Where next, GMOs?
Cell Biology in Europe - Portugal
BSCB Autumn Meeting Report