Integrins

Integrins and integrin ligands

Structural organisation of integrins

Integrins are heterodimeric receptors with a single transmembrane domain per monomer, able to bind the components of the extracellular matrix (ECM) through their extracellular part. With 18 distinct a subunits (genes ITGA) and 8 β subunits (genes ITGB), a total of 24 different integrins can be assembled (Fig. 10.1). The extracellular part of integrins is considerably larger than their intracellular part; it can be represented as an N-terminal globular head above two legs.

Textbook of Cell Signalling in Cancer_ An Educational Approach-Springer International Publishing (2015) 10.1

Fig. 10.1. Integrin assembly from their a and β subunits. The 18 a chains and 8 β chains can assemble in diverse ways to form the 24 integrins. Some integrins are specifically expressed in leukocytes and serve as immunoglobulin receptors. Some a chains have an interaction domain with the extracellular matrix (ECM); others do not and signalling operates only with the β -chain interaction (β I) domain

The globular head of the a chains consists of seven identical domains arranged as the blades of a propeller. Some integrins contain in addition, upstream the propeller domain, a domain analogous to type A von Willebrand factor (aI integrins). This domain enables the fixation of a divalent cation, Ca2+ or Mg2+, which interacts with the ECM; it is called the metal ion-dependent adhesion site (MIDAS). Downstream the propeller domain, the a chains contain domains sequentially called thigh, genu, calf1 and calf2, then the transmembrane domain and the intracellular domain (Fig. 10.2a).

Textbook of Cell Signalling in Cancer_ An Educational Approach-Springer International Publishing (2015) 10.2

Fig. 10.2. Integrin structure and spatial conformation of a and β chains. (a) Schematic representation of the characteristic domains of the a and β chains of integrins along the polypeptidic sequence. The aI domain of interaction with ECM is not present on all a chains. (b) Spatial representation of integrins. The propeller domains of the a chains are organised in a globular part, the head; the β I domains are exposed following the folding of the hybrid domains over the PSI domains. (a) Integrins adopt a folded conformation in the absence of stimulation; (b) under the effect of an intracellular stimulation (inside-out signalling), integrins adopt an extended conformation; (c) they can thus unmask a binding site with the extracellular matrix (ECM), which allows the interaction between the cytoplasmic domains and the generation of a signal (outside-in signalling)

The β chains also contain a globular head, a β I domain analogous to the aI domain but constant, bearing a MIDAS that follows a folding of the N-terminal extremity; this domain is inserted between two homologous hybrid domains, themselves inserted between two homologous PSI (plexins, semaphorins and integrins) domains. Downstream are four EGF domains, a tail domain (TD), the transmembrane domain and the intracytoplasmic part (Fig. 10.2a).

The intracytoplasmic domains of a and β chains are short but are functionally important; the interaction between these two cytoplasmic domains is stabilised by an ionic bond between an arginine residue of the a chain and an aspartic acid residue of the β chain.

In the absence of stimulation, integrins present a bent conformation, with the globular head lying down on the membrane (Fig. 10.2b). Intracellular messages induce their straightening, allowing them to interact with extracellular ligands. Ligand– integrin binding takes place at the level of the aI domain, for those having such a domain, and at the level of the β I domain for the others. These domains exist under two conformations, one closed, with low affinity for the ligand, and the other one open, with a MIDAS bound to Ca2+ or Mg2+, enabling the formation of a ionic bond with an aspartic acid residue of the ligand (Fig. 10.2b). Binding of integrins to their ligands induces a conformational change of the cytoplasmic domains of the a and β chains. The a chains are constitutively phosphorylated, whereas the β chains display phosphorylation sites allowing the regulation of their activity. Diverse protein kinase C (PKC) enzymes are involved in the phosphorylation of β chains serine/threonine residues, but it appears that tyrosine residues may also be phosphorylated.

An important feature in inside-out signal transduction is the aggregation of several integrin dimers; such a clustering is required in order to increase the attachment of the cell to the ECM. All along the polypeptidic chains of the matrix proteins are present numerous integrin binding sites, requiring a sufficient number of cell adhesion molecules for efficient gripping; these clusters of integrin dimers attached to the ECM are called focal adhesions. Signal transduction arising out of these clusters depends on both affinity changes (conformational switch of the intracytoplasmic domain) and avidity changes (clustering of integrin dimers).

Integrin ligands

The main extracellular ligands of integrins, considered as receptors, are the ECM proteins: fibrinogen (FG), fibronectin (FN), von Willebrand factor (VWF), collagens (COL), laminin (LAM) and vitronectin (VTN). Some integrins display high ligand specificity, while others are in contrast of low specificity. Particularly, a tripeptidic sequence present in some ECM proteins is specifically recognised by aV integrins: the RGD (arginine–glycine–aspartic acid) sequence. This sequence can serve as a guide for the design of competitive inhibitors of integrins; some snake venoms take advantage of this sequence for the disruption of the ECM. Other integrins recognise different tripeptidic sequences, such as IET or LDV, and some integrins recognise more complex tridimensional structures.

A subgroup of integrins (see Fig. 10.1) are selectively present at the surface of leukocytes, where they participate to immune functions. These integrins, such as aLβ 2 or LFA1 (lymphocyte function-associated antigen 1) and aMβ 2 or MAC1 (from macrophage), can recognise molecules of the immunoglobulin superfamily, such as ICAMs (intercellular adhesion molecules) and VCAM (vascular cell adhesion molecule). The five ICAMs (ICAM1 to ICAM5) are leukocyte transmembrane immunoglobulins whose binding to integrins allows cell–cell adhesion (see Chap. 11). Other extracellular integrin ligands are bacterial polysaccharides and viral proteins.

On the intracellular side, integrins are able to interact with many proteins involved in cytoskeleton structure regulation; these proteins bind the integrin β chain on the one hand and actin on the other hand, linking thus the ECM to the cytoskeleton. The most important of these proteins are talins (TLN), kindlins (FERMT, for fermitin family homologues), a-actinins (ACTN), filamins (FLN), cytohesins (CYTH), parvins (PARV) and tensins (TNS). These proteins not only are structural proteins, but they display diverse functions which are involved in transduction of signals received and transmitted by integrins. Adapter proteins are also involved in signal transduction: vinculin (VCL), paxillin (PXN) and various 14-3-3 proteins. The layout of integrin clusters with these proteins constitutes the focal adhesions (Fig. 10.3).

Textbook of Cell Signalling in Cancer_ An Educational Approach-Springer International Publishing (2015) 10.3

Fig. 10.3. Integrin-activated signalling pathways. A focal adhesion is schematised in the middle of the graph, showing the association between extracellular matrix (ECM) and intracellular actomyosin cytoskeleton (ACT/MY), via the assembly between integrins and several interactants (parvin, paxillin, cofilin, talin, actinin), some of which only are represented here. Two signalling pathways generated by integrin activation are shown: On the left, ILK (integrin-linked kinase), recruited, for instance, by PI3 kinase, phosphorylates the integrin β chains as well as several focal adhesion proteins; ILK also activates exchange factors (GEF) of small G-proteins of the RHO family, such as CDC42 and RAC1, which, once activated by GTP binding, can activate in turn some kinases like PAK, which can directly act on cytoskeleton proteins. On the right, FAK (focal adhesion kinase), recruited through integrin activation, phosphorylates focal adhesion proteins as well as SRC family kinases (SFK). Adapter proteins are able to activate proliferation pathways such as the AKT and the MAP kinases pathway. FAK also activates, via GEFs, small G-proteins of the RHO family, which, once activated, can activate in turn some kinases like ROCK (RHO-activated kinase), which can act on cytoskeleton proteins, either directly or through the phosphatase MLCP (myosin light chain phosphatase)

Signalling pathways arising from integrins

Integrins operate via two signalling ways: inside-out signalling, allowing intracellular messages to induce extracellular ligand recognition and binding via integrin activation, and outside-in signalling, as all receptors, allowing extracellular messages to induce specific cell responses. Integrins are therefore bidirectional receptors. These two activities are however inseparable: intracellular signals originating especially from TKR or GPCR activation activate first the receptor function of integrins by inducing their straightening, which governs ligand recognition; afterwards, this binding with extracellular ligands activates the formation of focal adhesions clusters and generates various intracellular messages. Integrin activation leads especially, according to cell type and integrin dimer, to the activation of cytoplasmic kinases, among which the tyrosine kinase FAK (focal adhesion kinase) (gene PTK2, protein tyrosine kinase 2) and the serine/threonine kinase ILK (integrin-linked kinase). These kinases are committed to various signalling pathways, leading either to the activation of transcription factors or to the activation of small G-proteins. These pathways are presented on Fig. 10.3.

FAK contains an N-terminal FERM (four-point-one, ezrin, radixin and moesin) domain, a central domain bearing the tyrosine kinase activity and a C-terminal FAT (focal adhesion targeting) domain, allowing FAK binding to paxillin and talin, and consequently its recruitment to focal adhesions. Integrin-mediated FAK activation operates through autophosphorylation, leading to the recruitment of other cytoplasmic tyrosine kinases of the SRC family (SFK, SRC family kinases), which complete FAK phosphorylation and are phosphorylated by FAK in return. FAK can also phosphorylate other proteins at the level of focal adhesions, such as a-actinin and paxillin.

FAK bears domains recognised by the SH2and SH3-binding sites of various proteins:

  • Via its SH3 domain-interacting sites, FAK induces the activation of GDP–GTP exchange proteins (GEF) of small G-proteins of the RHO family, such as p190RHOGEF or RGNEF (RHO-guanine nucleotide exchange factor), which are involved in cell motility through the control of the actin cytoskeleton.
  • Via its phosphotyrosine residues recognised by proteins with SH2 domains, FAK activates adapter proteins such as GRB2, the classical activator of the MAP kinase proliferation pathway (Chap. 2); p85, the regulatory subunit of PI3 kinase (Chap. 3), also governing proliferation and survival pathways; NCK (noncatalytic region of tyrosine kinase); CRK (chicken tumour virus regulator of kinase) and its associate CAS (CRK-associated substrate), which are involved in cell motility; and finally the SHC (SH2 domain-containing) proteins, which are involved in multiple cell processes.

ILK bears an N-terminal domain with ankyrin (ANK) repeat sites, allowing binding to PINCH (particularly interesting new Cys–His-rich) proteins, equipped with LIM (LIN-11, ISL-1 and MEC-3) domains and also known as LIMS (LIM and senescent cell antigen-like domains) proteins, to a phosphatase called ILKAP (ILKassociated protein phosphatase) and to integrin β chains. ILK also bears a central PH (pleckstrin homology) domain, allowing its activation by phosphatidylinositol 3,4,5-trisphosphate (see Chap. 3), and a C-terminal domain of binding with parvins (PARV), which are themselves actin-binding proteins and constitute a complex with ILK and PINCH, known as IPP (ILK, PINCH and PARV). ILK belongs to focal adhesions and is responsible for the phosphorylation of its partners, the integrin β chain, parvins, cofilins (CFL), paxillin, etc.

Beyond the focal adhesions, ILK main substrates are the serine/threonine kinases AKT (Chap. 3) and GSK3β (Chaps. 7 and 9), both involved in cell survival and proliferation. ILK also activates small G-proteins of the RHO family, such as RHOA, RHOD, CDC42 (cell division cycle 42), RND1 and RAC1 (RASrelated C3 botulinum toxin substrate 1), involved in cell motility as mentioned above.

Downstream the two integrin-activated kinases, FAK and ILK, one can find other kinases involved in cell proliferation and survival as well as the activation systems of small G-proteins of the RHO family, involved in cell motility. These proteins interact with the actomyosin cytoskeleton and induce the modifications required for cell migration. They promote actin polymerisation and the assembly of polymerised filaments, allowing thus the formation of filipods and lamellipods involved in cell movements. CDC42 and RAC1 activate various proteins such as WASP (Wiskott– Aldrich syndrome protein) and PAK (p21-activated kinase). RHOA induces the assembly and contraction of actomyosin fibres, especially through the activation of ROCK (RHO-associated kinase), which inhibits MLCP (myosin light chain phosphatase).

There exist multiple connections between the integrin pathway and the pathways induced by tyrosine kinase receptors (TKR, Chap. 1), especially their common activation of the MAP kinase pathway (Chap. 2) and the PI3 kinase pathway (Chap. 3), via the integrin-dependent kinases, FAK and ILK. NCK appears as an important link between the two signalling pathways; this adapter protein bears SH2 and SH3 domains. The integrin pathway is as important for the control of cell proliferation and survival as for the control of cell adhesion and motility; the cytoplasmic kinases FAK and ILK and the small G-proteins of the RHO family are the preferred mediators of these effects.

Some integrins are able to induce apoptosis when they are not bound to their ligand. They behave as dependence receptors (Chap. 18); this mechanism is called IMD (integrin-mediated death) and operates through caspase 8 activation. It especially concerns integrin aVβ 3, a cell proliferation-associated integrin. This integrinmediated death mechanism can be inhibited through the recruitment of cytoplasmic tyrosine kinase of the SRC family (SFK), which induce cell proliferation and survival via the activation of appropriate transcription factors.

Oncogenic alterations

Various integrins are overexpressed in cancers, such as a6β 4, aVβ 3, aVβ 5, a5β 1 and a4β 1, and they appear to play a major role in metastatic tumour development. These integrins cooperate with growth factors of the EGF and PDGF families (Chap. 1) and exert a positive effect on cell proliferation. Other integrins are in contrast negatively regulated in cancers, such as a2β 1, which activates the p38 MAP kinase pathway (Chap. 2), or aVβ 6 and aVβ 8, which contribute to TGFβ activation (Chap. 5); these integrins exert, therefore, a downregulation of cell proliferation.

Because of their effects on cell migration, integrins are central to the epithelialto-mesenchymal transition, which allows tumour cells of epithelial origin to acquire a mesenchymal phenotype capable to migrate and disseminate. These are again aVβ 6 and aVβ 8 integrins that can favour tumour invasiveness through TGFβ activation. Germline mutations of some integrin genes are found in some systemic diseases such as epidermolysis bullosa (integrin a6β 4) and congenital muscular dystrophy (integrin a7β 1), but they have not been found in cancers.

Endothelial cells express integrins in response to proangiogenic factors secreted by tumour cells; the binding of these integrins to ECM proteins (collagens, fibronectin, vitronectin, etc.) controls proliferation and migration of these endothelial cells, contributing thus to vessel formation and maturation. The aV integrin chains can associate with various β chains; integrin aVβ 3, equipped with a RGD-binding domain present on fibronectin, vitronectin and fibrinogen, is more abundant in tumour vessels than in normal tissue vessels. It is overexpressed during wound healing and inflammation, suggesting that inflammation could well be at the origin of the angiogenic switch of tumours. Integrin aVβ 3 plays thus a major role in survival and migration of endothelial cells; its antagonists induce endothelial cells apoptosis in vitro as in vivo.

Several other integrins have a major effect on angiogenesis. Integrin aVβ 5, closely related to aVβ 3 and which binds vitronectin, is induced by VEGF. Integrin a5β 1, the ligand of which is fibronectin, is mainly induced by bFGF in endothelial cells. Integrin a4β 1, stimulated by VEGF and bFGF (FGF1), recognises VCAM (vascular cell adhesion molecule) as a ligand, which enables the attachment of endothelial cells and pericytes to the smooth muscle cells that express this cell adhesion molecule on their surface. Integrins a1β 1 and a2β 1 have opposite roles in angiogenesis; they recognise mainly the collagens and their expression in endothelial cells is controlled by VEGFA and VEGFC, which suggests a role in both blood and lymphatic vessels.

Pharmacological targets

Because of the major role played by integrins in cell adhesion and migration, they constitute a potential therapeutic target in oncology. Three approaches have been developed for integrin targeting: peptides and peptidomimetics that can imitate their ligand, monoclonal antibodies and small molecules interfering with their activity.

The RGD tripeptide is the docking place on the ECM for several integrins, especially aVβ 3; peptides that mimic this sequence are susceptible to block cell signalling events induced by integrin activation when they are attached to the matrix. Natural peptides originating in particular from snake venoms and are called disintegrins have been tested as anticancer agents, but synthetic peptides are preferred, such as cilengitide, which entered clinical trials but was shown to be devoid of activity.

Monoclonal antibodies directed against specific integrins have been developed outside the field of oncology, and this approach is promising. The assembly of the a and β chains raises the problem of the optimal strategy: should specific chains be targeted or rather the assembly of two chains in a definite integrin? In addition, the extended conformation of integrins seems a better target than its folded conformation. In oncology, several antibodies have entered clinical trials: etaracizumab (targeting aVβ 3), volociximab (a5β 1), intetumumab (aV subunit) and natalizumab (a4 subunit). Finally, small non-peptidic molecules able to interact with integrins at the level of their docking sites on ECM have been identified; they can mimic, for instance, the RGD structure of ECM proteins interacting with aVβ 3 and other integrins (GLPG-0187) or the interaction of the a4 subunit with VCAM (AJM-300) or the interaction of a5β 1 with fibronectin (JSM-6427).

Downstream integrin receptors, the FAK and ILK proteins represent bona fide targets of the integrin pathway. The FAK-encoding gene, PTK2, is overexpressed in several cancer types, especially during metastatic spreading, but no activating mutations have been identified. The mechanism of this overexpression remains unknown, but the fact that the PTK2 gene promoter harbours sites repressed by p53 and activated by NFkB suggests the existence of crosstalks between these transcription factors and FAK expression. FAK targeting is conceivable, as it displays tyrosine kinase activity: ATP-competitive inhibitors have entered clinical trials. The compounds in development display cross-reactivity with a closely related cytoplasmic tyrosine kinase, PYK2 (proline-rich tyrosine kinase 2) (gene PTK2B), and with IGF1R, a receptor of the insulin-like growth factor subfamily (Chap. 1).

The ILK protein is activated in cancers, especially when a mutational or transcriptional defect of PTEN is present, which usually constitutes a potent brake for all proliferation effects depending on phosphatidylinositol 3,4,5-trisphosphate. This activation is associated to ILK overexpression, and ILK activating mutations have not yet been identified in cancers. ILK inhibition by antisense approaches or by serine/threonine kinase inhibitors is under study, because it induces in vitro important effects on cell proliferation and migration, especially in cells harbouring an invalidating PTEN gene mutation. However, in view of the redundant character of the signalling pathways activated by ILK, one can doubt about the therapeutic efficiency of ILK targeting.

0

Добавить комментарий

Войти с помощью: 

Ваш e-mail не будет опубликован. Обязательные поля помечены *