Extracellular Matrix

How the matrix is built

The ECM is not a passive packing material. It is a precisely engineered, dynamic, three-dimensional scaffold that mechanically supports tissues, transmits forces, presents adhesion ligands to cells, and stores growth factors for controlled release. It is also under constant remodeling — synthesized by fibroblasts, chondrocytes, and osteoblasts; degraded by matrix metalloproteinases; and re-synthesized after injury.

The four main classes of ECM molecules and how they fit together:

1. Fibrillar collagens (types I, II, III, V, XI). Synthesized as procollagen, hydroxylated on prolines and lysines (vitamin-C-dependent), assembled into right-handed triple helices, secreted, cleaved by procollagen N- and C-proteinases, and self-assembled into striated fibrils that are then crosslinked by lysyl oxidase. The ~67 nm "D-period" repeat is the textbook EM fingerprint.
2. Network-forming collagens (IV, VIII, X) and elastin. Type IV is the basement-membrane scaffold. Elastin, deposited on fibrillin-1 microfibrils and crosslinked by lysyl oxidase, gives skin, lung, aorta, and ligaments their recoil — a single elastin fiber can stretch and rebound a billion times in a lifetime.
3. Adhesive glycoproteins (fibronectin, laminin, tenascin, vitronectin). These present cell-binding motifs (RGD in fibronectin; LG domains in laminin) that integrins recognize. Laminin trimers tile the basement membrane; fibronectin assembles into fibrils under cell traction.
4. Proteoglycans (aggrecan, perlecan, decorin, syndecans, glypicans). Core proteins with covalently attached glycosaminoglycan chains — heparan, chondroitin, keratan, dermatan sulfate. The dense negative charge attracts water; the chains also bind growth factors and modulate their gradients.

Cells read this matrix through integrins, syndecans, dystroglycan, and CD44. Integrin engagement clusters the receptor, recruits talin and kindlin to the cytoplasmic tail, links to actin via vinculin and α-actinin, and triggers FAK, Src, MAPK, PI3K, and Rho-family signaling. Matrix mechanics — stiffness, fiber alignment, ligand density — are translated into transcriptional programs through YAP/TAZ.

Why the ECM matters

• Mechanical support. Bone (mineralized collagen I), cartilage (collagen II + aggrecan), tendon (parallel collagen I bundles), skin (collagen + elastin), and aorta (elastin lamellae) are all ECM-defined tissues.
• Cell signaling. Matrix stiffness alone can direct stem-cell differentiation: soft like brain → neurons, intermediate like muscle → myocytes, stiff like collagenous bone → osteoblasts (Engler et al., 2006).
• Growth-factor reservoir. Heparan sulfate proteoglycans bind FGF, Wnt, BMP, hedgehog, VEGF — the matrix functions as a regulated extracellular pharmacy.
• Wound healing. Provisional fibrin-fibronectin matrix is replaced by collagen III, then remodeled into mature collagen I scar over months.
• Fibrosis. Excess deposition of collagen and crosslinks underlies pulmonary fibrosis, cirrhosis, scleroderma, and cardiac fibrosis — a major cause of organ failure.
• Cancer. ECM stiffening promotes tumor invasion; matrix metalloproteinases enable metastasis; ECM-derived pre-metastatic niches dictate where circulating tumor cells implant.
• Tissue engineering. Decellularized ECM scaffolds (heart, kidney, liver) are used as templates for organ regeneration; collagen and hyaluronan dermal fillers are a multi-billion-dollar cosmetic industry.

Major ECM components

Component	Class	Where	Mechanical role	Disease link
Type I collagen	Fibrillar collagen	Skin, bone, tendon, dentin	High tensile strength (~100 MPa)	Osteogenesis imperfecta (COL1A1/2); classical EDS
Type II collagen	Fibrillar collagen	Cartilage, vitreous humor	Compressive resistance with aggrecan	Stickler syndrome; SEDC
Type IV collagen	Network collagen	Basement membrane	Sheet-like 2D scaffold	Alport syndrome (COL4A3/4/5); Goodpasture
Elastin	Elastic fiber	Skin, lung, aorta, ligamentum nuchae	Reversible stretch; recoil	Cutis laxa; supravalvular aortic stenosis (Williams)
Fibrillin-1	Microfibril scaffold	Aorta, suspensory ligaments, skin	Template for elastin assembly; sequesters TGF-β	Marfan syndrome (FBN1)
Laminin	Adhesive glycoprotein	Basement membrane	Cell-anchorage sheet; integrin and dystroglycan ligand	Congenital muscular dystrophy (LAMA2); junctional epidermolysis bullosa
Fibronectin	Adhesive glycoprotein	Provisional matrix; interstitial	RGD-bearing scaffold; assembled by cell traction	Glomerulopathy with fibronectin deposits
Aggrecan	Proteoglycan (chondroitin/keratan sulfate)	Cartilage	Holds 50× its weight in water; springy compression	Spondyloepiphyseal dysplasia; osteoarthritis
Heparan sulfate proteoglycans (perlecan, syndecans, glypicans)	Proteoglycan	Basement membrane; cell surface	Growth-factor binding; co-receptors	Schwartz-Jampel syndrome; Simpson-Golabi-Behmel (GPC3)
Hyaluronan	Non-sulfated GAG (no core protein)	Loose connective tissue, ECM, vitreous	Hydration; lubrication; signals via CD44	HA accumulation in tumors; cosmetic dermal fillers

Forms of ECM

Form	Tissue location	Main components	Stiffness range
Basement membrane	Beneath every epithelium and endothelium; around muscle, fat, Schwann cells	Laminin, type IV collagen, perlecan, nidogen	Soft sheet (~kPa)
Interstitial matrix	Loose connective tissue, dermis	Collagen I/III, fibronectin, hyaluronan, proteoglycans	~1-10 kPa
Cartilage matrix	Articular cartilage, growth plate	Collagen II, aggrecan, hyaluronan	~0.5-1 MPa compression
Bone matrix	Cortical and trabecular bone	Collagen I + hydroxyapatite mineral	~10-20 GPa
Tendon/ligament	Musculoskeletal connections	Parallel collagen I bundles, decorin	~1-2 GPa tensile
Elastic tissue	Aorta, lung, skin	Elastin on fibrillin-1 microfibrils	Low modulus, high strain
Tumor stroma	Cancer-associated	Crosslinked collagen I, fibronectin EDA, tenascin	Stiffened to 5-10 kPa+ in carcinomas

Real-world consequences

• Scurvy. Vitamin C is the cofactor for prolyl- and lysyl-hydroxylases that stabilize the collagen triple helix. Without it, defective collagen is made, blood vessels and connective tissues fail, and James Lind's 1747 citrus trial works.
• Osteogenesis imperfecta. Mutations in COL1A1 or COL1A2 produce defective type-I collagen. Glycine substitutions in the Gly-X-Y repeat distort the triple helix; bones break under normal loads.
• Marfan syndrome. FBN1 mutations weaken the fibrillin-1 microfibril scaffold; the aorta dilates and dissects; the lens dislocates; height increases. Fibrillin also sequesters latent TGF-β, so loss-of-function mutations release excess TGF-β signaling — a separate disease driver.
• Pulmonary fibrosis. Idiopathic pulmonary fibrosis (IPF) progressively replaces alveolar matrix with collagen-rich scar; the lungs stiffen, gas exchange fails, median survival is three to five years. Pirfenidone and nintedanib slow progression.
• Tumor stiffness. Breast tumors are 3-10× stiffer than normal breast (5-10 kPa vs. 150 Pa). The ECM stiffening is felt by integrins, activating YAP/TAZ and driving proliferation. Lysyl oxidase inhibitors are in trials as anti-fibrotic adjuvants.
• Stem-cell mechanotransduction. Engler, Sen, Sweeney, and Discher showed in 2006 that mesenchymal stem cells differentiate purely on the basis of substrate stiffness — soft polyacrylamide → neural lineage; medium → myogenic; stiff → osteogenic. Mechanics is a developmental signal.
• Decellularized scaffolds. Whole-organ decellularization with detergents leaves an intact ECM scaffold of laminin, collagen, and proteoglycans. Recellularizing such scaffolds with patient-derived cells is the central strategy for organ engineering.

Variants and special cases

• Collagen subfamilies. Twenty-eight types in humans. Beyond fibrillar (I, II, III, V, XI) and network-forming (IV, VIII, X), there are FACITs (IX, XII, XIV — fibril-associated), beaded (VI), anchoring fibrils (VII, in skin), and transmembrane (XIII, XVII, XXIII, XXV).
• Mechanically active fibronectin. Cell-applied tension stretches fibronectin fibrils and exposes cryptic binding sites — including a self-association site that drives fibril extension. Fibronectin is a force sensor as well as a substrate.
• Provisional matrix in wound healing. Within minutes of injury, plasma fibrin and fibronectin form a temporary clot-matrix. Fibroblasts migrate in along this scaffold and replace it with collagen III, then collagen I.
• Matricellular proteins. SPARC, thrombospondin, osteopontin, tenascin, and CCN family proteins are non-structural ECM components that modulate cell-matrix interactions, growth factor activity, and morphogenesis without bearing significant load.
• ECM in development. Heparan sulfate proteoglycans set up morphogen gradients (Wnt, BMP, FGF) that pattern embryos. Loss of HSPG biosynthesis is embryonic lethal.
• Plant cell wall. Often called the plant ECM equivalent — cellulose microfibrils, hemicellulose, pectin, and structural proteins (extensins) — but evolutionarily and chemically unrelated to animal ECM.

Common pitfalls and misconceptions

• "The ECM is just packing material." It is a signaling scaffold, a growth-factor reservoir, and a force sensor. Cells read its composition and stiffness as actively as they read soluble hormones.
• "Collagen and gelatin are the same." Gelatin is denatured (unwound) collagen — the triple helix is gone. Mechanical and signaling properties of native collagen depend on the triple helix and crosslinking.
• "Fibrosis is just too much collagen." Pathological fibrosis involves altered crosslinking, fiber orientation, and incorporation of fibronectin EDA isoforms — qualitative as well as quantitative changes.
• "Stiff matrix is always bad." Bone needs to be stiff. Cartilage needs to be soft. The pathology is the wrong mechanics for the tissue context — fibrotic stiffening of soft tissue or osteoporotic softening of bone.
• "Integrins are the only matrix receptors." Syndecans, dystroglycan, CD44 (hyaluronan), DDR1/2 (collagen), and LAIR-1 also read matrix; integrins dominate but do not have a monopoly.
• "MMPs only degrade matrix." They also cleave growth factors, growth-factor receptors, and chemokines, releasing or activating signaling molecules — they are proteases of the secretome more broadly.
• "Collagen oral supplements rebuild your skin." Ingested collagen is digested into amino acids and dipeptides like any other protein. Skin collagen synthesis depends on amino-acid availability and vitamin C, not on whether the dietary protein was branded "collagen."