From Structure to Phenotype: Impact of Collagen Alterations on Human Health

1The extracellular matrix (ECM) is a non-cellular complex network that provides a structural scaffold to the surrounding cells and, at the same time, a deposit of cytokines and growth factors capable of influencing cell behaviour.

2The main ECM components include collagens, proteoglycans and glycoproteins. In addition, many proteins such as growth factors, cytokines and proteolytic enzymes are associated to the ECM. Through all these components the ECM provides environmental information to the cells, which in turn respond by adapting their behaviour and adjusting proliferation, migration and differentiation. The ECM is highly dynamic and its remodelling has to be tightly regulated in order to maintain tissue homeostasis. Deregulation of the ECM structure or composition contributes to the onset of a variety of pathological conditions characterized by a wide range of tissue alterations, further straightening the functional relevance of the ECM. After a general overview of the ECM main features and functions we will focus on the most abundant elements of the ECM, the collagens. We will discuss the different collagen types, their synthesis and structural organization as well as the relevance of properly assembled collagen fibres by discussing their impact on human health.

3The ECM is a highly dynamic and heterogeneous structure. Each tissue has an ECM with a unique composition generated in early embryonic stages and then maintained and remodelled throughout the entire life. The great complexity of ECM structure and functions makes the research in this field very challenging. A significant input derives from the definition of the “matrisome,” a list of proteins contributing to the ECM in different organisms and tissues, which has been predicted by integrating the information derived from experimental knowledge, genome comparative studies and bioinformatics tools. This plastic and evolving list of ECM proteins, comprising between 1% and 1.5% of the mammalian proteome, requires further investigations and biochemical characterization.

4Hundreds of ECM components are known, many of which are capable of binding to other ECM components on specific sites, thus making the matrix a highly intricate structure (Figure 1).

5Although the molecular composition can vary widely, the ECM main components are: proteoglycans, hyaluronan, adhesive glycoproteins (fibronectins and laminins) and fibrous proteins (collagens and elastin). Proteoglycans are constituted by large carbohydrates (generically referred as glycosaminoglycans, GAGs) attached to a protein core. The anionic polysaccharides GAGs allow the sequestration of water and other cations such as calcium. Several types of protein core exist and various types of GAGs can bind to each other in a wide variety of combinations. Proteoglycans promote cell-ECM adhesion but also bind to secreted proteins and growth factors in the ECM. Many components of this category have space-filling and lubrication functions. A special type of GAG is hyaluronan (HA), the only GAG element of the ECM that lacks a protein core. It is a linear polysaccharide composed of the repeating disaccharide units N-acetyl-d-glucosamine and d-glucuronate. It is particularly abundant in human tissues, including joints, eyes, umbilical cord, synovial fluids, skeletal tissues, hurt and lung. In spite of its simple structure, HA plays a role in several biological processes, including cell signalling, inflammation, wound healing and cell development and contributes to maintain tissue homeostasis by interacting with other ECM proteins or proteoglycans. Thanks to its hygroscopic features, it acts as space filler among cells and contributes to the maintenance of tissue hydration. Its biocompatibility makes HA suitable for tissue engineering and clinical applications, where it can be used as a diagnostic marker.

6In mammals, around 200 glycoproteins provide interactions with other ECM components thus allowing ECM formation and assembly. They share multiple repeating domains and motifs typical of the ECM constituents and promote cell adhesion, cell signalling and binding to growth factors. The best-studied ECM glycoproteins are fibronectins and laminins. Fibronectins are proteins encoded by a single gene through multiple alternative splicing. The dimerization of two fibronectin monomers results in the final molecule, which contains repeating units named type I, II and III. Type I and II are tight together by disulphide bonds, the third one has seven-stranded β-barrel composition. All these modules are organized in order to contain binding sites for a variety of other molecules, such as heparin sulphate proteoglycans (HSPGs), integrins and collagens. Fibronectins can be soluble or associated to fibrils thus acting as bridging factors among different ECM components and anchoring cells to matrix fibres. Laminins are large cross-shaped ECM proteins composed of α, β and γ chains. In particular, twelve genes encode 5α, 4β and 3γ chains, which can differentially combine to generate many types of laminins. Among these, laminin G domain-like (LG) of the laminin α2 chain is composed by a β sandwich structure and contains a calcium-binding site, surrounded by a large number of epitopes involved in the interaction with cellular receptors and extracellular ligands ( and references therein).

7Fibrous proteins include collagens and elastin. Collagens are the most abundant proteins of the ECM and a detailed description of their structure and functions is found below (Section 4). Elastin is a key element of the ECM that provides elasticity and flexibility to different tissues including large arteries, ligaments, tendon, lung, skin and cartilage. It is synthesized and secreted as tropoelastin, a soluble precursor implicated in the formation of elastic fibres through the interaction with the N-terminal domains of fibrillins 1 and 2. Tropoelastin contains several hydrophobic domains (consisting in proline, glycine, valine and alanine) that are responsible for the extensibility properties of the protein. The tensile strength provided by collagen fibrils is therefore counterbalanced by the extensibility of elastic fibres, which, conversely to collagen, can undergo progressive stretching and relaxation cycles. Dynamic tissues are thus able to sustain mechanical stress without being permanently affected but reverting the tissues back to their original shape, a property known as viscoelasticity. The ECM tensile strength is determined by the dynamic activities of lysil oxidase (LOX) and lysil hydroxylase, two enzymes involved in regulating the cross-linking between collagens and elastin.

8Although not structural components of the matrix, many growth factors can bind to elements of the ECM and therefore are categorized as ECM constituents. Indeed, it was shown that vascular endothelial growth factor (VEGF) binds to the type III modules of fibronectin and this interaction depends on the heparin-binding residues of fibronectin. VEGF and fibroblast growth factor (FGF) can bind to HSPGs from where they are cleaved off as soluble ligands by the heparanase enzyme. Hepatocyte growth factor (HGF) binds to the 70 kDa N-terminal and the 40 kDa C-terminal fragments of fibronectin, whereas platelet-derived growth factor (PDGF) binds the type III and the variable domains of fibronectin. Differently, transforming growth factor-beta (TGFβ) binds to the latent transforming growth factor beta binding protein (LTBPs), which in turn binds to fibrillins and fibronectin-rich matrices. Also, fibrillin-containing microfibrils of the ECM regulate the availability and activity of bone morphogenetic proteins (BMPs) and growth and differentiation factor-5 (GDF-5), cytokines of the TGFβ family. Overall, the different growth factors associated to the ECM could be released locally and become available for the interaction with their canonical receptors. Thus, the ECM serves as a storage for growth factors and chemokines, whose interactions with the matrix control their half-life, local concentration and biological activity.

9For many years, the ECM has been defined as a static structure whose unique function was to provide support and shape to cells and tissues. Although this passive role is definitively fundamental for organism organization and maintenance, it is now clear that the ECM is much more than simple scaffolding. Synthesized and organized by the cells, the matrix itself can actively regulate cell behaviour. Indeed, the ECM provides a substrate over which cells can adhere and migrate by sensing the ECM constituents. In turn, cells will secrete new elements that result in ECM remodelling. Therefore, through the coordinated action of its main molecular constituents, the ECM can influence cell proliferation, adhesion and migration as well as differentiation and cell death.

10The ECM plays important structural roles during development and in particular during the formation of the skeleton. In tissues with mechanical functions such as cartilage, bone and tendons, the ECM is the major component that confers structural properties.

11Thanks to its biologically diverse array of macromolecules, the ECM provides a robust and dynamic scaffold capable to evolve during the normal physiological development but also to face insults that could disrupt tissue homeostasis. ECM reactions to these conditions are mediated by its physical, biochemical and biomechanical properties. Depending on its chemical composition, its topography and dimensionality, the ECM exerts different stimuli on the cells, which sense these forces and in turn respond to them. Cell migration, which is critical for proper normal embryonic development, is one of the best examples. Both motile (like the immune cells) and non-motile cells (such as adult epithelial cells) sense the composition and density of the surroundings and respond by migrating towards or moving away from the source.

12Tissue elasticity also depends on the ECM chemical composition, which defines soft or stiff matrices. Human tumours for example are surrounded by a stiff matrix with high collagen concentrations and the ECM rigidity might represent an optimal growing milieu for some surrounding cells that are therefore attracted towards this source. This phenomenon, known as desmoplasia, is usually associated with malignant tumours and it is present in many solid tumours. Tissue stiffness can drive malignant transformation via integrin-mediated mechanisms and such fibrotic “stiff” lesions are associated with a poor prognosis.

13ECM protein receptors, including integrin, syndecan, discoidin domain receptors (DDRs) and proteoglycans, can act synergistically to anchor the cells to the matrix and favour the reciprocal matrix organization. These adhesion dynamics are particularly important to maintain the right balance between self-renewal and differentiation of stem cells. In this respect, it has been shown that the genomic loss of integrin β1 encoding gene in the basal cells of mouse mammary epithelium affects stem cell regeneration and results in irregular branching ducts due to developmental defects of the mammary gland.

14As previously mentioned, the ECM can sequester several growth factors that are not structural components of the matrix per se but become active elements of the ECM. Chemokines, cytokines and growth factors, such as VEGFs, Wnts and FGFs, can be retained by the ECM and therefore create a “reservoir” of signalling molecules. By retaining these factors, the ECM may preserve the ligand source in proximity of the receiving cells and prevent their diffusion to the extracellular space. In addition, through the adhesion with ECM components, the ECM can modulate the ligand-receptors interaction and control the formation of morphogen gradients, whose concentration regulates developmental processes. One example of ECM molecule directly implicated in the regulation of morphogens gradients is represented by the HSPGs, which binds morphogens but also many cell surface co-receptors, thus acting as a linking platform that mediates the interactions of morphogens with the other ECM components.

15The ECM also contributes to ligand maturation. One representative example is the TGFβ proteins, which indirectly connect to fibrillins and fibronectins and are stored in the ECM in their inactive form until proteolitically activated by matrix metalloproteinases (MMP) or by mechanical forces. Furthermore, the ECM can trigger signalling events, as shown by the biologically active fragments derived from the proteolytic cleavage of collagens, proteoglycans, elastin and laminins ( and references therein). Matrikines were first described by Maquart and colleagues in 1999 and then named matricryptins one year later by Davis and colleagues with the following definition “biologically active sites that are not exposed in the mature, secreted form of ECM molecules but which become exposed after structural or conformational alterations”. Among them, endostatin derived from collagen XVIII is the most extensively studied matricryptin. The ectodomains of membrane collagens XIII, XVII, XXIII and XXV are matricryptins involved in cell adhesion, migration or proliferation. Matrikines from collagens IV are involved in angiogenesis and synapse formation. The ectodomains of syndecans 1–4 are also matricryptins, whereas fragments derived from hyaluronan degradation regulate inflammation and wound healing ( and references therein). Therefore, these ECM-fragments may act by regulating cell proliferation, cell death, cell differentiation and angiogenesis. Also, ECM receptors play a role in signal transduction, in particular the collagen receptors DDRs with their intracellular tyrosine kinase activity and the integrin proteins capable of transmitting chemical signals into the cells. Upon ligand binding, receptors get activated and trigger intracellular signalling events that through the involvement of Rho, Rock and the pathway of MAP kinases modulate cellular survival, proliferation and differentiation.

16The ECM provides several different contributions to the developmental events where it plays a dual role, both as functional as well as structural supporting element. A well-known example is provided by the morphogens, soluble factors contributing to define the patterning of surrounding cells during embryonic development. Morphogens are produced in restricted areas of the embryo from which they diffuse and, thanks to the presence of the ECM, generate gradients of signalling molecules that influence cell migration, adhesion and contractility by the activation of intracellular signalling pathways. Meanwhile, the ECM behaves as a structural element by defining the roads for cell migration, delimiting differentiating tissues and maintaining the shape of developing organs. In this context, the ECM assumes architectural roles, such as insulation of tissue to avoid nonspecific adhesion between tissues or, conversely, mediating adhesion between different tissue layers. The opposite functions of insulating or gluing embryonic tissues together show the flexibility of the ECM, which can select one or the other or even synchronize both events according to environmental stimuli. An example of coexistence is observed during muscle differentiation, where a sticky ECM is required to bind the extremities of the cells, whereas a slippery matrix coats the lateral sides. A slippery ECM is defined as an intact basement membrane (composed by laminin, integrin and glycoproteins) that allows tissues to freely slide on each other. GAGs with their negative charges and the resulting chain-chain repulsions, are the main cause of tissue slippery. Conversely, the removal of laminin induces a fragmentation of the basement membrane that, together with the presence of cell adhesion molecules (CAM), results in tissue sticking.

17ECM plasticity is also fundamental during branching morphogenesis of several vertebrate organs such as lung, kidney and mammary gland but also in skeletal development. During osteogenesis, skeletal progenitor cells undergo several morphological changes to ultimately give rise to the adult bone. In mature bone, the ECM is the result of active and opposing remodelling events exerted by the osteoclasts and osteoblasts, which degrade and deposit the bone matrix, respectively. An imbalance between degradation and deposition leads to alteration of bone density and disease.

18The ability of cells to move is central for embryo development as well as maintenance of multicellular organisms. Cell movement strictly depends on the balance between adherence to and release from the ECM in a dynamic fashion. The adhesive properties of the cells are mainly regulated by integrins, which play both structural and signalling roles. Integrins can sense the physical state of the matrix, by interacting with specific ECM molecules, such as collagens and laminins and activate downstream intracellular signalling cascades involving the focal adhesion kinase (FAK) signalling pathway, the mitogen-activated protein (MAP) kinases and the Rho family GTPases. Cells tend to migrate along oriented fibrils in a non-random movement. The removal of specific ECM components at a specific time, such as MMP-dependent proteolysis, can instead reorganize the ECM structure and therefore alter or promote the migration process. Thus, the ECM is not only a substrate but it plays dynamic and opposing roles in regulating cell migration. On one side, the basement membrane with its dense fibrillar protein network acts as a barrier to migrating cells, on the other, the ECM promotes cell movement by exposing chemotactic factors that can attract or repulse cells. The ECM remodelling contributes to the formation of organized pathways along which the cells can migrate in an oriented way. Collective cell migration along oriented patterns is an essential aspect of wound healing, a multi-steps process in which the skin repairs itself after injury. During this process, the ECM regulates the interactions between epidermal, dermal and bone marrow cells, it influences cell proliferation and orchestrates the deposition of new connective tissue and the migration of keratinocytes to the wound site.

19To fulfil its activities, the ECM requires a constant and regulated remodelling whose precise orchestration is crucial for tissue homeostasis and developmental processes characterized by transient and dynamic signalling events. ECM remodelling implies changes in ECM composition (novel synthesis or degradation of specific ECM components) or ECM architecture (modification of the macromolecule organization). Several enzymes are involved in ECM remodelling, below is provided a short description of the most known and characterized enzymes.

20Matrix metalloproteinases (MMPs) are a large family of enzymes that participate in the degradation of all major ECM components, including those of the basement membrane. They are zinc-dependent endopeptidases initially secreted in the extracellular environment as inactive zymogens with a pro-peptide domain that needs to be removed to allow enzyme activation. The MMPs can be either soluble or membrane-bound and present a substrate-specificity. At least 24 MMPs proteins have been so far identified and, based on their structural organization and substrate specificity, MMPs can be classified into: collagenases, gelatinases, stromelysins, matrilysins and membrane-type I. Under physiological conditions MMP activities are tightly regulated but they may increase during pathological events.

21Adamlysins, also called ADAMs (a disintegrin and metalloproteinases) and ADAMTS (ADAMs with a thrombospondin motif), are ECM proteinases involved in cell phenotype regulation, adhesion and migration. The ADAMs include both transmembrane and secreted proteins, whereas the ADAMTSs only contain secreted proteins. 21 ADAMs and 19 ADAMTS are known. They share several structural domains including the metalloproteinase as well as the disintegrin domain, the latest being involved in the binding to integrins. ADAMs are involved in cytokines processing and growth factor receptor shedding while ADAMTS play a role in degradation of ECM components, particularly collagens and proteoglycans.

22Meprins are membrane-bound or secreted metalloproteinases capable to cleave ECM molecules including the collagen type IV and fibronectins. In addition, they are involved in the synthesis of mature collagen molecules and in the activation of other metalloproteinases including MMPs and ADAMs.

23The right balance between ECM degradation and deposition has to be guaranteed for the correct tissue integrity. It is therefore evident the relevance of ECM proteinase inhibitors. The tissue inhibitors of metalloproteinases (TIMPs) represent a small family composed of only four members with the function of reversibly inhibiting the activity of MMPs and ADAMs. TIMPs present two distinct domains, one at the N- and one at the C-terminal region of the protein, which are responsible for the binding and inhibition of MMP activity, respectively. Although each TIMP molecule is active against various MMPs, they all show some substrate preferences.

24Other enzymes may be involved in ECM remodelling. Various proteolytic enzymes, including serine proteinases, cathepsins, heparanases, sulphatases and hyaluronidases, were shown to target ECM proteins. Plasmin and elastase are serine proteinases, the first degrades fibrin, fibronectin and laminin, whereas the latter degrades fibronectin and elastin. Cathepsins are lysosomal proteases, which are appointed to the degradation of intracellular or endocytosed proteins. Under specific circumstances cathepsins can be secreted in the extracellular environment where they contribute to the ECM protein degradation. Heparanases and sulphatases cleave heparin sulphate and remove its 6-O-sulphate residues, respectively. They affect the ability of heparin sulphate to bind several growth factors such as VEGF, PDGF and FGF, thus altering the downstream signalling events. Hyaluronidases are a family of enzymes capable of degrading HA.

25Collagens are the major insoluble fibrous proteins in humans and other vertebrates, accounting for about a quarter of their total protein mass. So far 28 different types of collagens have been identified in vertebrates. They assemble to adopt a triple-helix conformation that gives rise to long thin fibrils or two-dimensional reticulum or even associate with other ECM elements. The different types of collagens and their structure are crucial to provide mechanical stability, elasticity and strength to tissues and organs.

26Fibrillar collagens are the most abundant collagens in humans and they are synthetized as long precursors, known as procollagens, which contain a large polypeptide extension at both the N- and C-terminal ends. The C-propeptide has an essential role inside the rough endoplasmic (ER) reticulum where it initiates the assembly of three coiled subunits (α chains) one around the other and along a central axis in order to generate right-handed triple-helix. In vertebrates, over 40 genes encode collagen α chains, which are differentially combined to form 28 different collagen types. Despite the different structural organization, all collagen types share the triple-helix structure. An essential element for the assembly of the three α chains is the proline-rich tripeptide Gly-X-Y repetition, which characterizes all collagens. In the triple helix, glycine residues are localised in the central part, thus allowing a close packing of the molecule. Proline and hydroxyproline residues usually occupy the X and Y positions of the tripeptide. Moreover, hydroxylation of prolines and lysines in the middle region of the chains allows the formation of intra-molecular hydrogen bonds that stabilize the entire complex. The extent of lysine hydroxylation varies between tissues and collagen types. Some of the hydroxylysines are further modified by glycosylation with galactose and glucose. Notably, the short N- and C-terminal portion of the chains, which do not assemble in the triple-helix, are required for the extracellular secretion of the polypeptide and the formation of collagen fibrils. The N- and C-propeptides (telopeptides) are subsequently removed by procollagen aminoproteinases and procollagen carboxyproteinases, respectively, giving rise to tropocollagen units. Finally, adjacent tropocollagens are bound together through the formation of intermolecular interactions that involve lysine and hydroxylysine residues, thus providing the tensile strength of collagen fibrils. Finally, fibrils assemble into fibres of larger diameter. Details on the synthesis of the other collagen types are included (Section 4.2) and shown in Figure 2.

27Collagens can be grouped based on their structure, function and tissue distribution. They are designated by Roman numerals according to the order of their discovery (I-XXVIII). They are formed by three identical chains (homotrimers) or by two/three different chains (heterotrimers). The most abundant collagen of the human body, the interstitial type I collagen, is made by two identical α1 and one α2 chain, which shows high sequence homology with α1. In most other cases, including the collagen type II, they are homotrimers made by three identical α1 chains. The length of the triple helical region differs among the various collagens. The tripeptide repetition is the predominant motif in fibril collagens whereas it is much shorter and frequently interrupted by non-triple helical domains in other collagen types (such as the non-fibrillar collagens). Non-collagenous (NC) regions may also have structural function as shown by the transmembrane collagens.

28Following a classification based on collagen function and composition, we can distinguish:

29Fiber-forming collagens: they are characterized by a fibrillar shape and a rope structure. Under electron microscopy they show characteristic banding pattern. Fibril collagens assemble to form fibres whose diameter ranges from 12 to >500 nm and the length varies depending on the tissue and developmental stage. They are stabilized by non-reducible covalent crosslinks among specific triple-helix domains and telopeptides. They include the most abundant collagens of the organisms, such as the interstitial collagens (types I, II and III) and the collagen types V and XI, whose main functions consist in providing structural support, balance of pulling forces and enabling cell movement.

30FACITs (fibril-associated collagens with interrupted triple helices): they contain short collagenous regions with interruptions in the triple helix intercalated by four NC regions. These molecules are mostly heterotrimers and carry a glycosaminoglycan side chain. They include collagen types IX, XII and XIV, which associate with various collagen fibrils.

31Network-forming collagens: they are non-fibrillar collagens that aggregate linearly or laterally to form open networks. They are longer than classical collagens and can give rise to different kinds of networks depending on the collagen type. In particular, collagen type IV, the main component of epithelial basement membranes as well as vascular basal lamina, is irregularly assembled. Differently, collagen types VIII and X form regular hexagonal networks. The function of these network-forming collagens varies and likely depends on their structural organization.

32Transmembrane collagens: they are expressed in many different tissues and cells. This group of collagens plays an important role in epithelial and neural cell adhesion as well as in epithelial–mesenchymal interaction during morphogenesis. They are characterized by the presence of several triple helical regions in the extracellular C-terminal domain interspersed by NC stretches. Next to the extracellular portion of the protein, there is a conserved coiled-coil domain essential for the trimerization of transmembrane collagens. Collagen types XIII and XVII are included in this group.

33MULTIPLEXINs (multiple triple-helix domains and interruptions): collagen types XV and XVIII consist of several collagen domains with NC interruptions in the triple helixes, which are able to form oligomeric assemblies. They are found in some basement membranes covalently linked to glycosaminoglycan chains. The NC1 domain of collagen types XV and XVIII includes a peptide (endostatin) that following proteolytic cleavage is released in the extracellular environment. Several studies show the anti-angiogenic properties of endostatin as inhibitor of endothelial cell migration and tumour growth.

34Anchoring fibrils: collagen type VII is the major component of the anchoring fibrils, whose function is to secure the adhesion of the epidermal and dermal layers. It consists of a central collagenous triple-helical domain flanked by NC1 and NC2 domains. The NC2 domain is proteolytically cleaved while NC1 is preserved to anchor other ECM elements, including collagens and laminins. The anchoring filaments are assembled in an antiparallel manner, tail to tail with some C-terminal overlap.

35Beaded-filament-forming collagen: collagen type VI is the archetypal beaded filament-forming collagen. It is widely expressed and holds up tissue integrity. Collagen VI monomers are made up of short triple helical domains, which aggregate linearly to form beaded filaments or laterally through their globular domains, thus creating 3D networks. For this reason collagen type VI can also be included among the network-forming collagens. The N and C non-collagenous regions of the monomers are preserved and antiparallel dimers and tetramers are assembled intracellularly.

36Notably, collagen-like triple helical domains are found in several other proteins that do not have structural function and therefore are not considered as real collagens.

37Collagens have a great structural stability, resulting in high resistance against degradation by bacterial collagenases and other peptidases. Nevertheless, under physiological conditions most connective tissues undergo to a persistent turnover and continuous remodelling. Collagen degradation is a multi-step process that relies first on the activity of extracellular proteases to break down the ECM collagen fibrils and subsequently on the cellular uptake and intracellular lysosomal degradation of fragmented fibrils. The extracellular fragmentation of collagens is mainly mediated by proteinases such as the MMPs (collagenases and stromelysin), cysteine cathepsins and serine proteinases (plasmin). MMPs can target a wide range of ECM proteins, not only collagens. They act at neutral pH and recognize specific cleavage sites on the target molecules. Cathepsins are lysosomal proteases active at acidic pH, which can be active both intracellularly and upon secretion. Cathepsin S was shown to target and degrade collagens. An indirect collagen digestion can be achieved in response to plasminogen activation to plasmin. The pro-MMP-2 enzyme is activated by plasmin into MMP-2, also known as gelatinase A. Upon activation, MMP-2 can degrade several collagen types, fibronectin, elastin as well as gelatin, the denatured form of collagen. The extracellular collagen fragments are then recruited through phagocytosis from the neighbouring cells, mainly fibroblasts and macrophages, which send them to degradation via the lysosomal pathway. The relationship between the extracellular and intracellular pathways is complex and not fully understood. It has been shown that in macrophages uPARAP/Endo180 acts as collagen internalization receptor after the interaction with pro-uPA (pro-urokinase plasminogen activator) and uPAR (urokinase plasminogen activator receptor) proteins. Moreover, collagen internalization requires the expression of specific integrins and cytokines, including TGFβ and interleukin 1α. Finally, lysosomes fuse together to generate large structures containing collagen and ECM fragments that undergo enzymatic digestion by cysteine cathepsins.

38The tight regulation of ECM synthesis and remodelling is fundamental for human health, as attested by the high number of hereditary disorders caused by mutations in genes encoding structural elements of the ECM or proteases implicated in the remodelling process. Alterations of ECM remodelling can also influence the course and progression of several other pathological conditions, including fibrosis, skin disorders and cancer. The excessive ECM production and the concomitant loss of degradation leading to fibrosis will not be discussed in this review, which is focused on the human genetic disorders associated to an altered collagen structure (Table 1) or resulting in excessive degradation of specific collagen elements (Table 2). A general overview of the clinical features associated to these collagen-related disorders is also provided (Table 1 and Table 2).

39Since collagens are present throughout the entire body, alterations impairing the quality or quantity of collagen structures can affect any tissue or organ. Since each collagen is generally expressed in several different tissues and it is tightly associated with other ECM elements, alterations result in widely overlapping features that make the diagnosis difficult. Nevertheless, we thereafter propose a tentative classification of the collagen-related disorders according to the major clinical features and affected organs listed in Table 1 and Table 2.

40Skeletal and cartilage abnormalities: collagen type I is the major ECM component secreted by osteoblasts during bone development. Therefore, alterations in this molecule can give rise to the osteogenesis imperfecta (OI) characterized by bone fragility, as well as the Caffey disease with its infantile episodes of excessive new bone formation (hyperostosis). Also, mutations in COL2A1 gene result in skeletal abnormalities including the incomplete bone ossification in patients with achondrogenesis type II before birth or the short stature (dwarfism) of patients with Kniest dysplasia. Dwarfism can also be caused by alterations of collagen type IX or X (multiple epiphyseal dysplasia or the Schmid-type metaphyseal chondrodysplasia, respectively) or by mutations in COL11A1, COL11A2 or COL27A1 genes. Moreover, collagen types II, IX and XI are implicated in the formation and maintenance of the cartilage, thus they are relevant for the joint health and long bone development. Mutations in COL2A1 gene may result in cartilage alterations characterized by progressive degeneration at the joints (patients with osteoarthritis with mild chondrodysplasia), by hypercellular cartilage with large chondrocytes (Torrance type of platyspondylic lethal skeletal dysplasia) or by a translucent and abnormal gelatinous texture (achondrogenesis type II). Additionally, the presence of fibrous cartilage can be found associated with skeleton defects in patients with fibrochondrogenesis-1 or multiple epiphyseal dysplasia, due to mutations in COL11A or COL9A gene, respectively.

41Skin alterations: some of the collagen-related disorders present severe skin alterations. The dystrophic forms of epidermolysis bullosa (EB) with mutations in COL7A1 or COL17A1 gene is one of the major forms of EB where patients present a fragile skin, which can shed at the slightest touch. In milder cases blistering may affect the hands, feet, knees and elbows but in severe cases blistering may lead to vision loss, disfigurement and strictures of the gastrointestinal tract. Skin defects are also the main features of patients with the Ehlers-Danlos syndrome (EDS), a genetically heterogeneous disorder with more than nineteen causative genes. Mutations in COL5A1 or COL5A2 are responsible for the classical form of EDS. Patients present a soft, velvety skin that is highly stretchy (skin hyperextensibility) and fragile. Affected individuals tend to bruise easily and in some cases, they show atrophic scars. Skin alterations are also found in patients with Bethlem myopathy-1 caused by mutations in COL6A genes. Even though muscle dystrophy is the main clinical feature in this group of patients, they present follicular hyperkeratosis on the arms and legs, soft, velvety skin on the hand palms and feet soles, abnormal wound healing that leads to shallow scars.

42Finally, alterations of MMP1 gene expression have been associated to disorders with skin defects: overexpression of MMP1 in primary dermal fibroblasts of patients with trichothiodystrophy is responsible for collagen type I degradation and altered wound healing features whereas a functional single nucleotide polymorphism in MMP1 promoter is associated with increased collagen type VII degradation and high severity of recessive dystrophic EB.

43Hearing loss and visual defects: many of the collagen-related disorders are characterized by sensorineural hearing loss. In particular, all the disorders due to alterations of collagen type XI (see Table 1), the Alport syndrome and the X-linked deafness-6 due to alterations of collagen type IV and the OI and EDS disorders by collagen type I defects may reveal sensorineural hearing loss. Notably, some of these disorders are also associated to visual problems that may include myopia, cataract and in some cases (Stickler syndrome) retinal detachment. More severe vision defects can be observed in patients with the Knobloch syndrome-1 due to alterations of the type XVIII collagen. These patients are affected by high myopia, cataract, dislocated lens, vitreoretinal degeneration and retinal detachment. Finally, alterations of type IV collagen can also result in visual defects as observed in patients with retinal arterial tortuosity, Axenfeld-Rieger anomaly and Small vessel disease of the brain.

44Muscle weakness: collagen types VI certainly plays a relevant role in skeletal muscle maintenance and regeneration. Alterations of its major elements, the α1 and α2 chains, result in Bethlem myopathy-1, Ullrich congenital muscular dystrophy-1 or the autosomal recessive myosclerosis, all characterized by progressive muscle weakness (hypotonia) and joint stiffness (contractures) with different degree of severity. In the most severe cases weakness of respiratory muscles are reported. Differently, alterations affecting the α3 chain of collagen type VI are found in rare cases with dystonia 27. These patients reveal dystonic action and postural tremor mainly involving the face, neck, bulbar muscles and upper limbs. A severe generalized hypotonia leading to exercise intolerance, feeding difficulties and respiratory insufficiency are present in patients with congenital myasthenic syndrome type 19 due to mutations in COL13A1 gene. Also, patients with EDS reveal weak muscle tone and hypermobile joints, which can delay the development of motor skills such as sitting, standing and walking. Two cases with OI due to mutations in SPARC gene present underdeveloped muscles of the lower extremities, muscle hypotonia and gross motor developmental delay, whereas a single family with congenital fibrosis of extraocular muscles-5 caused by mutations in COL25A1 gene showed ophthalmoplegia of the extraocular muscles.

45Small vessel anomalies and kidney disease: the type IV collagen is the major constituent of the basement membranes. It is a non-fibrillar collagen made of three distinct heterotrimers generated by the products (α chains) of 6 distinct genes. Mutations in COL4A1 and COL4A2 result in thickness and damaged vascular basement membranes that affect the straightness of the vessels in patients with susceptibility to intracerebral haemorrhage, hereditary angiopathy or small vessel disease of the brain. Also, the retinal arterial tortuosity derives from mutations in the COL4A1 gene. Differently, alterations impairing the α3, α4 or α5 chains of collagen type IV result in defects of the glomerular basement membrane that affect kidney functionality. This is observed in patients with Alport syndrome who experience high levels of haematuria and proteinuria due to the progressive loss of kidney activity. Persistent and recurrent haematuria is also observed in the benign form of familial haematuria due to mutations in COL4A3 or COL4A4 genes.

46Most of the collagen-related disorders affect early childhood health but mild situations can also occur during adulthood. Indeed, the severity of clinical features in patients affected by Stickler syndrome varies among individuals and mild cases with late onset have also been reported. Conversely, when collagen alterations result in severe deformations, the survival of the entire organism is compromised as shown by PPIB mutations in type IX osteogenesis imperfecta. Patients affected by such severe form of OI die during gestation or shortly after birth. It is worthwhile considering that the severity of the disorder relies on several different factors where tissue distribution and function of the affected collagen play leading roles. The OI with its various degree of severity is the result of structural alterations of collagen type I, the most abundant collagen in humans. In particular, mutations inactivating one of COL1A1 alleles and resulting in reduced levels of an otherwise normal type I collagen are usually responsible for the mild forms of OI whereas dominant negative mutations in COL1A1 or COL1A2 genes account for the most severe forms. Notably, among the most common mutations responsible for the severe form of OI are those involving the substitutions of the glycine amino acid in the G-X-Y repeats, essential for the formation of the triple helix. This strongly points to the notion that structural alterations are more detrimental to human health than collagen impoverishment. Similarly, mutations in COL2A1 gene result in several rare autosomal dominant clinical entities that share skeletal dysplasia, short stature and sensorial defects. The wide range of clinical manifestations were not fully understood but a recent study on over 700 cases (harbouring 415 different mutations) revealed that one-third of the mutations affect the glycine amino acid in the G-X-Y repeats and give rise to the severe achondrogenesis type II disease, which is typically identified in utero and may result in embryo death. In contrast, mutations resulting in a premature stop codon or the p.Arg275Cys substitution are responsible for the less severe cases affected by Stickler syndrome or Czech dysplasia. In summary, the type of mutation, the tissue distribution and the function of the affected collagens all impact on the clinical spectrum of collagen-related disorders.

47The large number of genetic disorders associated to collagen alterations clearly strengthens the relevance of this wide group of proteins, which have been for long time considered inert elements with no other function than maintenance of tissue shape and architecture. The observation that mutations in collagen coding genes result in alterations of relevant developmental processes (skeletal and cartilage development) or defects of tissue homeostasis (skin, sensorineural, visual and muscle alterations) clearly demonstrate a regulatory role for this type of molecules. Therefore, not only the collagens but also the ECM with its broad number of elements and its wide complexity plays a role of primary importance among the mechanisms implicated in embryonic development, normal organ physiology and human health. Moreover, the wide and largely overlapping spectrum of clinical features of collagen, or even ECM-related disorders, makes in some instances the clinical diagnosis and patient management very difficult. A better understanding of the signalling events regulating the expression, functions and dynamic interplay of the various ECM elements will improve our knowledge on the pathogenesis of ECM-related disorders and, in parallel, will provide the tools for the identification of potential therapeutic targets.