Cell Adhesion Molecules

A cell adhesion molecule is a transmembrane (or GPI-anchored) glycoprotein whose extracellular domain binds a specific partner on another cell's surface, whose transmembrane domain crosses the plasma membrane, and whose cytoplasmic tail — for many CAMs — links to the cytoskeleton or signaling machinery. Binding can be homophilic, where a CAM on one cell binds an identical CAM on the apposing cell (typical of classical cadherins and many IgSF CAMs), or heterophilic, where it binds a structurally different molecule, often a carbohydrate rather than a protein (typical of selectins). CAMs also form cis interactions — lateral associations between identical or different CAMs within the same membrane — which cluster and strengthen the trans interactions that actually span the intercellular gap. Three families dominate direct cell–cell CAM biology, and each embodies a distinct chemical logic. Cadherins are calcium-dependent homophilic adhesion proteins built from tandem extracellular cadherin (EC) repeats; removing Ca²⁺ with a chelator like EGTA abolishes their binding almost instantly, a signature used experimentally to identify cadherin-dependent adhesion. Selectins are calcium-dependent but heterophilic: their N-terminal C-type lectin domain recognizes specific sialylated and fucosylated oligosaccharides (notably sialyl-Lewis^x) displayed on glycoprotein ligands, giving fast-forming, fast-breaking bonds suited to capturing cells from flowing blood. Immunoglobulin superfamily CAMs — NCAM, ICAM-1, ICAM-2, VCAM-1, PECAM-1 (CD31), L1-CAM, and others — are built from repeated Ig-like folds, bind calcium-independently, and can engage homophilically (NCAM–NCAM, PECAM–PECAM) or heterophilically (ICAM-1 binds the integrin LFA-1, VCAM-1 binds integrin VLA-4), linking the IgSF world directly to integrin-based adhesion covered in Integrins and Cell-Matrix Adhesion. The unifying principle across all three families is that adhesive specificity and strength are tunable properties, not fixed constants. Cells regulate CAM number, clustering, glycosylation state, cytoplasmic anchoring, and even the mechanical response of the bond itself (some bonds strengthen under tension — "catch bonds" — rather than weakening, as ordinary "slip bonds" do). This tunability is what allows the same basic molecular toolkit to build a stable epithelial sheet in one context and a transient, force-resistant leukocyte tether in another. Understanding CAMs is therefore a prerequisite for Tight and Anchoring Junctions, which describes how cadherin-based adhesion is elaborated into adherens junctions and desmosomes, and follows naturally from the general framework introduced in Cell Junctions Overview.

Cadherins: calcium-locked homophilic zippers. Classical cadherins (E-cadherin in epithelia, N-cadherin in neural and mesenchymal tissue, P-cadherin in placenta and skin) each have five tandem extracellular cadherin (EC1–EC5) repeats, a single transmembrane helix, and a cytoplasmic tail. Ca²⁺ ions bind at the linker regions between successive EC repeats, rigidifying the normally flexible chain into a curved, rod-like rod that resists bending. Adhesion begins when the EC1 domain of a cadherin on one cell inserts a conserved tryptophan side chain (Trp2) into a hydrophobic pocket in the EC1 domain of a cadherin on the opposing cell — a "strand-swap" interaction. This single trans bond is weak, but cadherins first associate laterally in cis within the same membrane, clustering several strand-swapped trans dimers together into an adhesive "zipper" that runs along the cell–cell contact and dramatically increases avidity. Intracellularly, the cadherin tail binds p120-catenin (which stabilizes the cadherin at the membrane and prevents its endocytosis) and β-catenin, which in turn binds α-catenin, linking the whole complex to the actin cytoskeleton — the structural basis of the adherens junction elaborated further in Tight and Anchoring Junctions. Because this link is dynamic, actomyosin contraction can pull on cadherin-based junctions, remodeling tissue shape during processes like gastrulation and wound closure.

Selectins: catch-bond tethers for rolling leukocytes. There are three selectins — L-selectin (on leukocytes), P-selectin (stored in endothelial Weibel-Palade bodies and platelet α-granules, rapidly surface-exposed after histamine or thrombin stimulation), and E-selectin (induced on endothelium hours after cytokine exposure, e.g., TNF-α or IL-1). Each has an N-terminal C-type lectin domain, an EGF-like domain, a variable number of complement-regulatory repeats, a transmembrane domain, and a short cytoplasmic tail. The lectin domain binds Ca²⁺-dependently to sialyl-Lewis^x and related tetrasaccharides carried on glycoprotein scaffolds such as PSGL-1 (P-selectin glycoprotein ligand-1). Selectin bonds have unusually fast on- and off-rates, which is exactly what is needed for a leukocyte moving at venous shear rates to make brief, repeated contacts with the vessel wall rather than sticking rigidly or not at all — the physical basis of "rolling" adhesion seen by intravital microscopy. Remarkably, selectin–ligand bonds are catch bonds: over a certain force range, increasing shear force increases bond lifetime rather than shortening it, because tensile force reorients the lectin domain into a higher-affinity conformation. This counterintuitive mechanical property, confirmed by flow-chamber and single-molecule force-spectroscopy experiments, keeps rolling leukocytes engaged specifically in flow conditions where adhesion is mechanically demanded, and lets bonds release again when shear falls below a threshold — a self-limiting braking system.

Immunoglobulin superfamily CAMs: calcium-independent bridges between adhesion systems. IgSF CAMs are built from stacked Ig-like β-sandwich domains (typically 2–7 repeats) that fold independently of calcium, so EGTA does not disrupt them — a classic way to distinguish IgSF-mediated from cadherin-mediated adhesion experimentally. NCAM (neural cell adhesion molecule) mediates homophilic binding between neurons and glia through its Ig domains; a subset of NCAM is modified with long chains of polysialic acid (PSA-NCAM), which sterically weakens adhesion and increases plasticity in regions of active neurite outgrowth and synaptic remodeling. ICAM-1 and ICAM-2, constitutively or inducibly expressed on endothelium, bind heterophilically to the integrin LFA-1 (αLβ2) on leukocytes; VCAM-1 binds integrin VLA-4 (α4β1). This ICAM/VCAM–integrin interaction converts the loose, reversible rolling initiated by selectins into firm arrest, the second step of the leukocyte adhesion cascade. PECAM-1 (CD31), expressed at endothelial cell–cell junctions, binds homophilically and guides the physical process of diapedesis — leukocyte squeezing between endothelial cells to exit the vessel. Because IgSF CAM binding is typically weaker per bond than cadherin binding but the domains support rapid clustering and signaling (through associated kinases and small GTPases), this family excels at transient, inducible, and highly regulated adhesion events rather than the stable, load-bearing adhesion cadherins provide in epithelia.

Townes and Holtfreter (1955): tissues sort themselves according to type, not randomly. Working with amphibian embryos, Philip Townes and Johannes Holtfreter dissociated the three germ layers — ectoderm, mesoderm, endoderm — into single cells using alkaline solutions, then mixed the cells together and allowed them to reaggregate. The cells did not remain randomly mixed. Instead, they sorted out, with ectoderm consistently ending up on the outside and endoderm on the inside, reconstituting a normal tissue architecture. This was the first rigorous demonstration that cells carry surface information specifying selective, type-appropriate adhesion — the conceptual founding observation for the entire CAM field, since it proved that adhesion is a specific, regulated recognition process rather than uniform stickiness.

Malcolm Steinberg (1960s): the differential adhesion hypothesis. Steinberg proposed and tested a quantitative model: if different cell types have different, measurable adhesive strengths, mixed populations should sort exactly as immiscible liquids of different surface tensions would — the more cohesive cell type moving to the interior, minimizing the total interfacial free energy. He tested this using cell types from chick embryos (e.g., heart, liver, neural retina cells) at controlled ratios and conditions, showing the sorting order matched predictions from measured aggregate cohesivity (assessed by how strongly aggregates resisted mechanical disruption). This differential adhesion hypothesis (DAH) explained Townes and Holtfreter's qualitative result quantitatively and correctly predicted sorting hierarchies in new combinations of cell types, cementing the idea that graded adhesive affinity, encoded by CAM type and level, is sufficient to explain tissue self-organization without any need for cells to "know" their destination.

Takeichi and colleagues (1977 onward): identifying cadherins as the molecular basis of Ca²⁺-dependent cell adhesion. Masatoshi Takeichi's laboratory showed that treating cells with EGTA (a calcium chelator) abolished aggregation of many cell types, revealing a class of Ca²⁺-dependent adhesion molecules distinct from Ca²⁺-independent ones. Using function-blocking monoclonal antibodies raised against surface proteins, they identified a specific glycoprotein whose blockade phenocopied EGTA treatment; they named this protein "cadherin" (calcium-dependent adherence). Subsequent molecular cloning revealed multiple cadherin genes (E-, N-, P-cadherin) expressed in tissue-specific patterns, and transfection experiments were decisive: L cells (a fibroblast line that is normally non-adhesive and does not express cadherins) became calcium-dependently adhesive and sorted from untransfected cells when transfected with E-cadherin cDNA alone. This gain-of-function experiment proved that a single cadherin gene product is sufficient to confer both adhesion and sorting specificity, directly linking a defined molecule to Steinberg's abstract "differential adhesiveness."

Springer, Butcher, and colleagues (1980s–1990s): dissecting the multistep leukocyte adhesion cascade. Intravital microscopy of postcapillary venules showed that leukocytes in inflamed tissue first roll slowly along the endothelium before suddenly arresting and then crossing the vessel wall — a sequence too fast and specific to be simple sticking. Function-blocking antibodies against selectins stopped rolling but not firm arrest, while antibodies against integrins (LFA-1) or their endothelial ligands (ICAM-1) stopped firm arrest but not rolling, showing that separate, sequential molecular systems govern each step. Selectin-knockout mice (L-selectin, P-selectin, E-selectin, singly and in combination) confirmed these roles in vivo: mice lacking selectins showed markedly reduced leukocyte rolling and delayed inflammatory responses, and combined selectin deficiency in mice and in the human disease "leukocyte adhesion deficiency type II" (caused by defective fucosylation needed to build selectin ligands) produces recurrent bacterial infections, directly linking selectin biochemistry to organismal immune competence. Parallel biophysical work — flow-chamber assays measuring rolling velocity as a function of shear stress, and later atomic force microscopy and biomembrane force probe experiments on single selectin–ligand bonds — established the counterintuitive catch-bond kinetics, in which bond lifetime initially increases with applied force, directly explaining why rolling is stable specifically under physiological shear and destabilizes if shear falls too low or rises too high.

Example 1 — Predicting sorting behavior from cadherin expression levels (a DAH-style calculation). Suppose three cultured cell populations express E-cadherin at very different surface densities, quantified by flow cytometry as molecules per cell:

Steinberg's differential adhesion hypothesis predicts that when all three types are mixed and allowed to reaggregate, sorting proceeds according to relative cohesiveness (which scales, to a first approximation, with adhesion molecule density and bond strength), such that the most cohesive population minimizes its surface exposure by moving to the core. Given the roughly 12-fold difference between A and C, the prediction is: type C cells (highest cadherin density) sort to the innermost core, type B forms an intermediate shell, and type A (lowest density, weakest self-cohesion) ends up on the outside, in direct contact with the culture medium. This is exactly the pattern later confirmed experimentally with L-cell transfectants expressing different levels of E-cadherin — cells expressing more cadherin molecules envelop cells expressing fewer, reproducing the liquid-like "engulfment" behavior predicted by minimizing interfacial energy. Note the key conceptual point: sorting here requires only a quantitative difference in the same homophilic adhesion molecule, not different molecules altogether — differential adhesion, not exclusive recognition, is sufficient.

Example 2 — Reasoning through a clinical adhesion defect: Leukocyte Adhesion Deficiency (LAD). A child presents with recurrent severe bacterial skin and mucosal infections, delayed umbilical cord separation, and, notably, wound sites that show little pus formation despite active infection — because neutrophils fail to accumulate at the site. Blood tests show markedly elevated circulating neutrophil counts (leukocytosis), since neutrophils are not being consumed/retained at infection sites. Two distinct molecular lesions can produce this picture, and distinguishing them requires knowing the adhesion cascade:

• LAD type I results from mutations in the gene encoding the β2 integrin subunit (CD18), which pairs with αL to form LFA-1. Without functional LFA-1, neutrophils roll normally on selectins (initial capture is intact) but cannot bind ICAM-1 and therefore never convert rolling into firm arrest — they are swept away in the bloodstream rather than adhering, and thus never reach the tissue to fight infection.
• LAD type II results from a defect in a fucose transporter needed to build fucosylated selectin ligands (like sialyl-Lewis^x). Here, rolling itself is defective from the start because selectins have nothing appropriate to bind, so neutrophils never make initial contact with inflamed endothelium at all.

Given the clinical description above (recurrent infection, poor pus formation, high circulating neutrophils), a clinician cannot distinguish LAD-I from LAD-II by symptoms alone; the diagnostic step is a flow-chamber rolling assay or flow cytometry for CD18 expression and sialyl-Lewis^x expression on the patient's neutrophils. If rolling is normal but firm adhesion fails, suspect LAD-I (test CD18 surface expression, typically near absent in severe LAD-I). If rolling itself is absent or grossly reduced, suspect LAD-II (test sialyl-Lewis^x expression and blood group, since the same fucosylation defect affects ABO/Lewis blood group antigens). This example shows how the two-step molecular mechanism (selectin-mediated rolling, then integrin/ICAM-mediated arrest) directly explains a two-way clinical differential diagnosis.

Masatoshi Takeichi trained in developmental biology in Japan and, in the late 1970s, became interested in why some cultured cells clump together tightly while others do not, and why this clumping was so sensitive to extracellular calcium. Rather than treating the calcium sensitivity as a biochemical curiosity, Takeichi pursued it as a clue to a specific molecule, using function-blocking antibodies systematically to isolate and name the responsible surface glycoprotein — cadherin. His insistence on molecular identification, followed by cloning and transfection experiments that showed a single cadherin gene was sufficient to confer both adhesion and sorting specificity to otherwise non-adhesive cells, transformed cadherins from a phenomenological curiosity into a defined molecular family with a clear structure–function logic. His later work extended the cadherin superfamily to include the protocadherins and clarified how tissue-specific cadherin expression patterns pattern the developing nervous system.

Cancer biology and metastasis. Loss of E-cadherin function — through mutation, promoter methylation, or transcriptional repression by factors like Snail and Slug during epithelial-to-mesenchymal transition — is one of the most consistent molecular events in invasive carcinoma, and germline E-cadherin (CDH1) mutations cause hereditary diffuse gastric cancer, prompting prophylactic gastrectomy in confirmed carriers. Restoring E-cadherin expression in cultured carcinoma cells experimentally suppresses invasiveness, directly demonstrating a causal (not merely correlative) role for cadherin loss in malignant behavior.

Anti-inflammatory and anti-thrombotic drug design. Because selectins initiate the leukocyte rolling that drives excessive inflammation in conditions like sickle cell vaso-occlusive crisis, selectin-blocking antibodies and small-molecule selectin antagonists have been developed and used clinically (e.g., crizanlizumab, a P-selectin-blocking antibody, reduces vaso-occlusive pain crises in sickle cell disease) — a direct translation of the rolling-adhesion mechanism into therapy.

Immunology and vaccine/adjuvant design. Integrin-ICAM interactions are targeted by drugs such as natalizumab (blocks VLA-4, reducing leukocyte entry into the CNS in multiple sclerosis) and efalizumab (blocked LFA-1, formerly used in psoriasis), showing how manipulating the "firm arrest" step of the adhesion cascade can control autoimmune and inflammatory disease, though such drugs carry infection risk precisely because they impair normal immune surveillance.

Neuroscience. NCAM and its polysialylated form regulate synaptic plasticity, and abnormalities in NCAM/PSA-NCAM levels have been implicated in schizophrenia and mood disorders, motivating ongoing research into adhesion-molecule-based biomarkers and therapeutic targets for neuropsychiatric disease.

Developmental biology and tissue engineering. Differential cadherin expression is used experimentally and increasingly in bioengineering to control which cell types sort together when building organoids and engineered tissues, directly applying Steinberg's differential adhesion logic to construct multi-layered tissue architectures in vitro.

Diagnostics. Flow-chamber rolling assays and flow cytometric panels for CD18, sialyl-Lewis^x, and specific cadherins are standard clinical and research tools for diagnosing leukocyte adhesion deficiencies and characterizing tumor invasiveness, translating the classic experimental assays described above directly into patient care.

Ask the tutor: How exactly does clustering of cadherins in cis strengthen adhesion beyond what a single trans bond provides, and why does this matter for tissue mechanics?