Cell Migration and Lamellipodia

Why cell migration matters

• It builds the embryo. Gastrulation, neural-crest streaming, and axon-growth-cone pathfinding are all migration. Neural-crest cells delaminate from the dorsal neural tube and crawl vast distances — to the face, the gut, the adrenal medulla, and the skin — using exactly this actin-and-adhesion machinery. Errors produce neurocristopathies such as Hirschsprung disease (an aganglionic colon where enteric neural-crest cells failed to arrive).
• It closes wounds. When you cut skin, epithelial cells at the wound edge lose their apical-basal polarity, extend lamellipodia, and crawl as a sheet to re-cover the gap — a process that takes hours to days and is imaged in the ubiquitous "scratch assay." Fibroblasts and myofibroblasts migrate into the clot to lay down and contract new matrix.
• It runs the immune system. Neutrophils are the fastest human cells, chemotaxing out of blood vessels toward bacterial fMLP peptides and complement C5a at over 10 µm per minute. Dendritic cells crawl to lymph nodes; T cells scan thousands of antigen-presenting cells per hour by crawling. Leukocyte-adhesion-deficiency (defective β2 integrins) causes recurrent infections precisely because these cells cannot grip and crawl out of the bloodstream.
• It kills in cancer. Metastasis causes roughly 90% of cancer deaths, and it is migration gone rogue: carcinoma cells undergo epithelial-to-mesenchymal transition, build matrix-digesting invadopodia, and crawl into blood and lymph vessels to seed distant organs.
• It is mechanically astonishing. A crawling cell is a soft machine that converts the chemical energy of actin polymerization and myosin ATP hydrolysis into directed motion with no wheels, no rigid skeleton, and no external anchor beyond a sticky floor — moving its own body weight forward against fluid drag and adhesion, and reversing on command when the gradient flips.
• It is a drug target. Anti-migratory and anti-invasive strategies — ROCK inhibitors, Rac inhibitors like EHT 1864, and MMP inhibitors — are being pursued to block metastasis and fibrosis, though the machinery's overlap with normal wound healing and immunity makes clean targeting hard.

How cell migration works, step by step

Mesenchymal crawling on a two-dimensional substrate runs as a four-phase cycle that repeats indefinitely: protrude, adhere, translocate, retract. The steps overlap in space along the front-to-back axis, so a moving cell is doing all four simultaneously in different regions.

1. Protrusion. At the leading edge, the small GTPase Rac1 (loaded with GTP by GEFs such as Tiam1 and DOCK180) together with the membrane lipid PIP2 activates the WAVE regulatory complex. WAVE presents an actin monomer and the flank of a "mother" filament to the Arp2/3 complex, a seven-subunit machine whose Arp2 and Arp3 subunits mimic an actin dimer. Arp2/3 nucleates a new "daughter" filament that branches off at a characteristic 70-degree angle. Because each daughter can be re-branched, the result is an exponentially expanding dendritic meshwork whose barbed (growing) ends all face the membrane. Capping protein terminates most filaments after ~100–200 nm, keeping them short and stiff; profilin-charged G-actin feeds the growing ends. The collective polymerization pushes the membrane outward by a Brownian-ratchet mechanism — thermal fluctuations of the membrane open a gap, a subunit inserts, and the filament ratchets the edge forward. In parallel, Cdc42 drives finger-like filopodia: parallel bundles of 10–30 unbranched filaments, cross-linked by fascin and elongated at their tips by formins (mDia2) and Ena/VASP proteins, that act as antennae probing the terrain.

2. Adhesion. Under the fresh protrusion, transmembrane integrins engage the extracellular matrix (fibronectin, collagen, laminin) and cluster into nascent adhesions. Adaptor proteins — talin, vinculin, paxillin, and the enzyme focal adhesion kinase (FAK) — assemble on the integrin tails and physically link them to the actin network. Talin acts as a mechanosensitive clutch: when actin retrograde flow pulls on it, talin unfolds and exposes cryptic vinculin-binding sites, reinforcing the adhesion under load.

3. Translocation. With the front anchored, the cell body and nucleus move forward. Actin filaments in the lamellipodium constantly flow rearward ("retrograde flow"); when the clutch engages the substrate, that flow is converted into forward traction rather than being wasted, and the whole cell is pulled up over its adhesions.

4. Retraction. At the rear, RhoA activates ROCK and the formin mDia1. ROCK phosphorylates myosin light chain (and inhibits its phosphatase), switching on myosin II. Bipolar myosin II filaments slide antiparallel actin filaments, contracting the network and generating the tension that hauls the trailing edge forward. Rear focal adhesions disassemble — microtubules target them for turnover, and the protease calpain cleaves talin and FAK to release the grip — so the tail detaches and snaps up. The cycle then repeats.

Directionality comes from chemotaxis. G-protein-coupled receptors read a soluble gradient; even a 1–2% concentration difference across a 10-µm cell is amplified internally into a steep compass by PI3-kinase making PIP3 at the front and the phosphatase PTEN clearing it at the sides and rear. This biases Rac1/Cdc42 activation — and therefore protrusion — toward the source (the local-excitation, global-inhibition model).

Lamellipodia vs filopodia vs invadopodia

Feature	Lamellipodium	Filopodium	Invadopodium
Shape	Broad flat sheet	Thin finger / spike	Ventral punctate protrusion
Actin architecture	Dendritic branched (Arp2/3)	Parallel bundle (fascin)	Branched + bundled core
Main GTPase	Rac1	Cdc42	Cdc42 / RhoC
Key elongators	Arp2/3, capping protein	Formin mDia2, Ena/VASP	N-WASP, cortactin, Tks5
Primary job	Generate protrusive force	Sense & steer	Digest the matrix (MT1-MMP)
Typical dimensions	1–5 µm deep, ~0.2 µm thick	0.1–0.3 µm wide, µm-long	~0.5 µm wide, µm-deep
Where it matters	General crawling, wound sheets	Growth cones, endothelial tips	Cancer invasion, metastasis

Mesenchymal vs amoeboid migration

Property	Mesenchymal	Amoeboid
Cell shape	Elongated, spread, front-rear polarized	Rounded, blebbing or gliding
Adhesion	Strong, integrin focal adhesions	Weak, transient or adhesion-independent
Leading structure	Arp2/3 lamellipodium	Membrane blebs, cortical flow
Contractility	Moderate (stress fibers)	High (RhoA/ROCK cortex)
Matrix degradation	Yes — MMP/invadopodia proteolysis	No — squeezes through pores
Speed	Slow (~0.1–1 µm/min)	Fast (up to ~20 µm/min)
Examples	Fibroblasts, epithelial sheets	Leukocytes, some carcinoma cells

Many tumor cells switch between the two modes ("mesenchymal-amoeboid transition") when one is blocked, which is why single-pathway anti-invasion drugs so often fail — inhibit the proteases and the cell simply rounds up and squeezes through instead.

Common misconceptions

• "The lamellipodium pulls the cell forward." It pushes the membrane outward by actin polymerization; the force is a compression against the membrane, not a pull. The forward pull on the cell body comes later, from myosin-II contraction at the rear and from traction where adhesions convert retrograde actin flow into substrate grip.
• "Faster actin polymerization means faster cells." Not usually. Barbed ends can elongate fast enough to move a cell far quicker than it actually goes; the rate-limiting step is coordinating adhesion turnover and rear retraction. Migration speed is biphasic with adhesion strength — too weak and there is no traction, too strong and the rear cannot release — so a cell peaks at an intermediate stickiness.
• "Cdc42 builds the lamellipodium." Cdc42 builds filopodia and sets overall direction/polarity. Rac1 builds the branched-actin lamellipodium via WAVE and Arp2/3. Confusing the two reverses the classic Nobes–Hall assignments.
• "Focal adhesions are static anchors." They are dynamic clutches in constant turnover. Nascent adhesions form under the protrusion and either disassemble within ~60 seconds or mature under tension; rear adhesions are actively dismantled by microtubules and calpain. A cell that could not disassemble adhesions would be glued in place.
• "Chemotaxis means the cell physically feels the gradient direction." The external gradient is often shallower than 2% across the cell — far too weak to read directly. The cell amplifies that tiny asymmetry internally through PIP3/PTEN signaling into a sharp intracellular polarity, essentially manufacturing a steep compass from a nearly flat cue.
• "All cells crawl the same way." Mesenchymal crawling on a dish is only one mode. In 3D tissue, leukocytes and some cancer cells use adhesion-independent amoeboid migration, and cells can migrate collectively as connected sheets or chains — with leader cells at the front and followers behind — rather than as lone crawlers.

Famous experiments and history

• Abercrombie's ruffling membrane (1960s–70s). Michael Abercrombie used interference-reflection microscopy on fibroblasts to describe the "ruffling membrane" at the leading edge and the phenomenon of contact inhibition of locomotion — cells stop and turn away when they touch a neighbor. His work established the front-to-back cycle of protrusion, attachment, and rear retraction that still frames the field.
• Nobes & Hall microinjection (1992–1995). Catherine Nobes and Alan Hall microinjected constitutively active Rho, Rac, and Cdc42 into fibroblasts and read out the cytoskeleton: activated Rac produced lamellipodia and ruffles, Cdc42 produced filopodia, and Rho produced stress fibers and focal adhesions. This defined the three-GTPase division of labor that underpins all modern models of migration.
• The fish keratocyte model. Epithelial keratocytes from fish skin crawl fast (5–10 µm/min) with a strikingly stable, fan-shaped lamellipodium that barely changes shape. Their steadiness made them the ideal system for measuring actin retrograde flow, the graded-radial-extension model of movement, and traction forces — Julie Theriot and colleagues used them to show the lamellipodial actin array treadmills as a coherent whole.
• Reconstitution of actin-based motility (1999). Marie-France Carlier, Thomas Pollard, and others reconstituted actin-comet-tail motility of the bacterium Listeria monocytogenes — and then of coated beads — from purified proteins (actin, Arp2/3, capping protein, ADF/cofilin, profilin, and a nucleation-promoting factor). This minimal system proved that Arp2/3-based branched-actin polymerization alone can generate propulsive force, the same engine that runs the lamellipodium.
• Traction-force microscopy. By plating cells on deformable elastic gels embedded with fluorescent beads and tracking bead displacement, researchers (Micah Dembo, Yu-li Wang, and others) mapped the piconewton-to-nanonewton traction stresses a crawling cell exerts — showing the strongest pull concentrates just behind the leading edge and at the retracting rear, exactly where the clutch engages and myosin contracts.