Focal Adhesions

Why focal adhesions matter

• The cell's mechanical sensor. Focal adhesions translate substrate stiffness into intracellular signaling. Mesenchymal stem cells on 1 kPa gels (brain-soft) become neurons; on 10 kPa (muscle-stiff) become myocytes; on 30 kPa (bone-stiff) become osteoblasts. The Engler, Sen, Discher 2006 Cell paper made this stiffness-driven differentiation a foundational result of mechanobiology — and the readout happens at the talin-vinculin layer of the adhesion.
• Required for cell migration. A migrating fibroblast assembles new adhesions at the leading-edge lamellipodium and disassembles them at the rear. The cycle takes 10 to 30 minutes per adhesion. Without functional integrins or talin, cells cannot generate traction; without disassembly, they cannot release the rear and move forward.
• FAK signaling is essential for embryogenesis. FAK knockout mice die at E8.5 with mesodermal defects. The kinase autophosphorylates at Y397 upon integrin clustering, recruits Src, and propagates signals through Ras-MAPK (proliferation), PI3K-AKT (survival), and Rac-Cdc42 (migration). FAK inhibitors like defactinib (VS-6063) are in clinical trials for mesothelioma, ovarian, and pancreatic cancer.
• Drug targets for thrombosis and cancer. Abciximab (ReoPro), the first FDA-approved monoclonal antibody, blocks platelet integrin αIIbβ3 and reduces ischemic events after coronary intervention. Tirofiban and eptifibatide are small-molecule αIIbβ3 antagonists in the same class. Cilengitide targeted αvβ3/αvβ5 in glioblastoma trials but failed phase III — a cautionary tale for integrin pharmacology.
• Adhesion stiffness sensing drives cancer progression. Tumor stroma stiffens up to 10x as collagen crosslinks accumulate. The stiff matrix activates focal-adhesion mechanotransduction in cancer cells, driving YAP/TAZ nuclear translocation, proliferation, and metastatic invasion. Lysyl oxidase inhibitors that prevent stromal stiffening are in oncology development.
• Fibrillar adhesions assemble fibronectin. Centrally located, elongated tensin-rich fibrillar adhesions translocate α5β1 integrin away from the leading edge along stress fibers, dragging fibronectin dimers and triggering them to unfold and polymerize into the fibrillar matrix that organizes the ECM. Without this, basement membranes fail to mature.
• Conserved across all metazoans. Sponges, jellyfish, flies, fish, and humans all use integrin-talin-vinculin-FAK adhesions with the same layered architecture. The core machinery predates the bilaterian split. Plants and fungi have independently evolved different ECM-attachment strategies because they lack integrins.

Common misconceptions

• All adhesions are focal adhesions. No — there is a developmental progression: nascent adhesions (~100 nm, <1 min, no actin connection), focal complexes (~1 µm, 1 to 10 min, lamellipodium/lamellum boundary), focal adhesions (1 to 5 µm, 5 to 30 min, anchor stress fibers), and fibrillar adhesions (elongated, central, organize fibronectin matrix). Each has distinct composition and dynamics.
• Integrins are always active. Integrins exist in three conformations: bent-closed (low affinity for ECM), extended-closed (intermediate), and extended-open (high affinity). Talin and kindlin binding to the β tail induces extension; ECM ligand binding stabilizes the open state. This bidirectional activation — inside-out and outside-in — is essential for regulated adhesion. Constitutively active integrins cause platelet hyperactivation and clotting.
• Focal adhesions are static. They are dynamic protein assemblies in continuous turnover. FRAP measurements show that even "mature" focal adhesions exchange their core components (vinculin, talin, paxillin) on timescales of seconds to minutes. The structure persists; the molecules cycle.
• Vinculin glues actin to talin. Vinculin is autoinhibited in the cytosol — its head and tail bind each other. Recruitment to talin's force-exposed binding sites separates head from tail and activates vinculin. Only then can it bind both talin (head) and actin (tail), serving as a force-induced molecular clutch. Static glue would not provide stiffness sensing.
• Higher force always means stronger adhesion. The catch-bond behavior of integrins means binding lifetime increases with force up to a peak, then decreases. Beyond ~30 to 50 pN per integrin–ligand pair, bonds rapidly break. Adhesions sustain large forces only by sharing load across many integrins (clustering).
• Adhesions form on any surface. Integrins recognize specific ligand motifs — the RGD tripeptide in fibronectin and vitronectin, GFOGER in collagen, LDV in fibronectin's CS-1, and so on. Polyethylene glycol surfaces are non-adhesive precisely because they lack integrin binding sites; cells round up and die (anoikis) if forced to remain.

How focal adhesions work

Adhesion biogenesis begins when a leading-edge protrusion contacts the ECM. Bent-closed integrin heterodimers (an α and a β chain together — humans have 18 α and 8 β subunits combining into 24 known αβ pairs) cluster underneath the membrane. Talin in the cytosol is autoinhibited (the FERM head buries the rod's vinculin-binding sites). PI(4,5)P2 enrichment at the membrane and Rap1-RIAM signaling break talin autoinhibition; the FERM head binds the integrin β cytoplasmic tail and induces extension — this is inside-out activation. The extended integrin then engages an ECM ligand (e.g., the RGD motif of fibronectin), completing outside-in activation. Kindlin co-binds the β tail at a distinct site and is also required for stable activation. Roughly 50 to 100 integrin–talin pairs cluster within ~100 nm to form a nascent adhesion.

The nanoscale architecture is stratified, not a soup. Pakorn Kanchanawong and Clare Waterman's 2010 super-resolution iPALM imaging of mouse embryonic fibroblasts revealed three vertical layers above the inner plasma membrane. The integrin signaling layer at ~10 nm contains paxillin and FAK, anchored to the integrin cytoplasmic tails. The force-transduction layer at ~30 nm is dominated by talin (oriented as an elongated rod, head-down) and kindlin. The actin-regulatory layer at ~50 to 60 nm holds vinculin, VASP, α-actinin, and zyxin, where contractile stress fibers attach. Talin's rod spans roughly the height between the integrin and actin layers and acts as the structural and mechanical bridge.

Mechanotransduction happens in the talin rod. Actomyosin contractility (myosin II generating ~5 pN per motor head, summed over many filaments) pulls on the actin layer; that load is transmitted through vinculin into the talin rod and out to the integrin. Force unfolds the helical bundles of talin one at a time, exposing eleven cryptic vinculin-binding sites. Vinculin binding both reinforces the connection (more actin attachment) and recruits more talin in a positive feedback. On stiff substrates the rod unfolds to high extents and the adhesion matures; on soft substrates the rod remains folded, vinculin doesn't bind, and the adhesion fails to mature and disassembles. The transcriptional readout converges on YAP/TAZ — co-activators retained in the cytosol when adhesions are weak, and translocated to the nucleus when adhesions mature, driving proliferation and survival programs.

Nascent vs focal complex vs focal adhesion vs fibrillar

Property	Nascent	Focal complex	Focal adhesion	Fibrillar
Size	~100 nm	~1 µm round	1–5 µm elongated	5–10 µm thin streaks
Lifetime	<1 min (~80% turn over)	1–10 min	5–30 min, sometimes hours	tens of minutes to hours
Location	Behind lamellipodium	Lamellipodium–lamellum boundary	End of stress fiber, cell body	Central, away from edge
Actin attachment	None	Forming	Anchors stress fiber	Linked, sliding
Key components	Integrin, paxillin, FAK	+ α-actinin, some vinculin	+ Vinculin, talin, zyxin	+ Tensin, α5β1 prominent
Force loading	Low (actin not engaged)	Increasing	High (myosin II contracts)	Translocating, lower than FA
Dependence on Rho-ROCK	Independent	Partial	Required (Y-27632 collapses)	Required for translocation
Function	Initial probe	Test traction	Sustained adhesion + signaling	Fibronectin matrix assembly

Integrin αβ pairs and their ligands

Integrin	Main ligand	Cell type	Disease association
α5β1	Fibronectin (RGD)	Fibroblasts, most cells	Tumor angiogenesis target
αvβ3	Vitronectin, fibrinogen, RGD	Endothelium, osteoclasts, tumor	Tumor angiogenesis (cilengitide)
αIIbβ3	Fibrinogen	Platelets only	Glanzmann thrombasthenia (loss); thrombosis (gain)
α6β4	Laminin-332	Basal keratinocytes (hemidesmosomes)	Epidermolysis bullosa pluriforme
αLβ2 (LFA-1)	ICAM-1, ICAM-2	Leukocytes	LAD type I (β2/CD18 mutations)
αMβ2 (Mac-1)	iC3b, fibrinogen	Macrophages, neutrophils	LAD type I
α2β1	Collagen (GFOGER)	Platelets, epithelium	Bleeding disorders (rare)
α4β7	MAdCAM-1	Gut-homing lymphocytes	IBD target (vedolizumab)

Famous experiments

• Abercrombie's interference reflection microscopy (1971). Michael Abercrombie used IRM at UCL to visualize the close contacts (later named focal adhesions) on the underside of crawling chick fibroblasts. The dark patches in IRM corresponded to membrane regions within ~10 nm of the substrate — the first evidence of discrete attachment sites rather than continuous adhesion.
• Kanchanawong & Waterman iPALM nanoscale architecture (2010). Using interferometric photoactivated localization microscopy on mouse embryonic fibroblasts, Pakorn Kanchanawong, Clare Waterman, and colleagues at NIH resolved focal adhesions into three distinct vertical layers spanning ~40 nm. Published in Nature 468: 580–584; redefined adhesions as architecturally precise nanostructures.
• Engler, Sen, Discher stiffness-driven differentiation (2006). Plating mesenchymal stem cells on polyacrylamide hydrogels of tunable stiffness, the Discher lab at Penn showed differentiation tracks substrate stiffness alone — neurons on 1 kPa, myocytes on 10 kPa, osteoblasts on 30 kPa — with no chemical induction. Cell 126: 677–689; foundational paper of mechanobiology.
• Talin-vinculin force unfolding (Yan & Sheetz labs, 2014). Single-molecule magnetic tweezers stretched individual talin rods and recorded force-induced unfolding of the helical bundles between 5 and 25 pN, with vinculin binding to the unfolded states. The experiment directly demonstrated talin as a tunable mechanical switch — the molecular basis of stiffness sensing.
• FAK knockout (Ilic et al., 1995). Generated FAK-null mouse embryos that died at E8.5 with mesodermal defects and disorganized somites. FAK-null fibroblasts showed enlarged, more numerous focal adhesions and reduced migration speed — paradoxical at first, but explained by the role of FAK in adhesion turnover, not formation.