Macrophage Phagocytosis and the CD47–SIRPα Checkpoint Explained

A macrophage does not decide to eat; it counts votes. Every particle it touches carries a mixture of pro-phagocytic and inhibitory ligands, and engulfment follows only when the activating tally clears a threshold set locally at the contact site. That arithmetic is why a healthy erythrocyte survives a splenic transit while an antibody-coated tumour cell does not, and why CD47 blockade became one of the most heavily pursued targets in innate immuno-oncology.

The eat-me signals and the receptors that read them

Three recognition systems feed one actin machine. Opsonic uptake begins with IgG: FcγRI (CD64), FcγRIII/II and the high-affinity mouse FcγRIV (CD16-2) all signal through the shared FcRγ ITAM chain, while FcγRIIB carries an ITIM and pulls the other way. The activating-to-inhibitory ratio on a given macrophage, rather than antibody affinity alone, predicts how well a therapeutic IgG drives antibody-dependent cellular phagocytosis.

The complement arm runs from C1q to C3b/iC3b, read by CR3 (CD11b) and CR4 (CD11c), both paired with the β₂ chain CD18 and both needing inside-out activation — supplied partly by C5aR1 (CD88) — before they bind. The apoptotic-cell arm reads externalised phosphatidylserine through the tether TIM-4, through MFG-E8 bridging to αvβ5, and through Gas6 and Protein S bridging to MerTK, Axl and Tyro3.

Two further inputs broaden the repertoire. Surface calreticulin engaging LRP1 (CD91) is a phosphatidylserine-independent eat-me signal that appears on stressed and transformed cells, and is much of the reason CD47 blockade spares normal tissue. The scavenger and lectin receptors MSR1 (SR-A), CD36, CD206, Dectin-1 and TREM2 add opsonin-independent pattern recognition.

From ITAM to actin: building the phagocytic cup

Receptor clustering brings Lyn to the FcRγ ITAMs; dual phosphorylation docks Syk, which recruits Btk, PLCγ2 and PI3Kδ. PI3Kδ generates PIP₃ in a sharply bounded membrane patch, and that lipid gradient — not the receptor itself — tells the cell where the cup forms. PTEN and SHIP-1 erase it at the rim, which is why SHIP-1 loss yields cups that fail to close cleanly.

Vav1 then activates Rac1 and Cdc42. Cdc42 dominates at the advancing edge, driving WASP-dependent nucleation through Arp2/3 to extend the pseudopod; Rac1 dominates at closure, reached by the phosphatidylserine route through a separate CrkII–DOCK180–ELMO1 GEF. Closure itself is contractile: RhoA and ROCK1 recruit Myosin-IIA to squeeze the cup shut. That step is precisely where the CD47 checkpoint acts, so a macrophage under SIRPα inhibition still binds and spreads — and any assay scoring conjugates rather than internalisation misses the phenotype.

The CD47–SIRPα checkpoint and its parallels

CD47 sits on essentially every nucleated cell and on erythrocytes, where its density falls with cell age. Ligation of macrophage SIRPα (CD172a) phosphorylates its ITIMs and recruits SHP-1 (PTPN6) and SHP-2 (PTPN11), which dephosphorylate the FcRγ ITAMs and the proximal signalosome while Csk holds Lyn in its inhibited conformation. The output is a raised threshold rather than an on/off switch.

CD47 is only one brake, and that redundancy is the practical problem. MHC-I/β₂-microglobulin engages LILRB1; sialylated CD24 engages Siglec-10/G; PD-L1 engages macrophage PD-1 and suppresses phagocytic capacity cell-intrinsically; CD200 engages CD200R; and homophilic CD31 (PECAM-1) contact actively repels viable cells. Tumours that lose CD47 can upregulate any of these, so single-axis blockade rarely gives a complete phenotype.

The therapeutic logic follows from the arithmetic: removing an inhibitory signal only helps if an activating one is already present, which is why CD47 or SIRPα blockade is nearly always combined with an opsonising IgG. It also explains the dose-limiting anaemia seen clinically, since ageing erythrocytes already carry natural IgG and sit just below threshold.

Phagosome maturation, the oxidative burst and LAP

The nascent phagosome is plain plasma membrane; everything degradative is acquired afterwards. Rab5 and EEA1 mark the early phase and drive fusion with early endosomes, and the Rab5-to-Rab7 conversion then licenses fusion with lysosomes, delivering LAMP1 and the acid protease cathepsin D. Tracking that conversion is the cleanest way to separate a degradative defect from a failure of uptake.

In parallel, cytosolic p47phox (NCF1) assembles onto the flavocytochrome NOX2 (CYBB) to generate the ROS burst. A distinct route, LC3-associated phagocytosis, conjugates LC3B directly onto the single-membrane phagosome using a PI3K complex containing Rubicon, which also stabilises NOX2. LAP speeds maturation but dampens antigen presentation — Rubicon-deficient mice develop lupus-like autoimmunity for that reason. Akt links uptake to macrophage survival.

The output: silent clearance versus inflammation

How a particle enters determines what the macrophage does next. TAM-receptor efferocytosis is deliberately quiet: MerTK signalling induces IL-10 and TGF-β and suppresses TNF and IL-1β, so billions of apoptotic cells are cleared daily without inflammation. FcγR- and Dectin-1-driven uptake does the opposite. Two experiments that both score "phagocytosis" can give opposite cytokine readouts, which is why the uptake assay and the supernatant ELISA belong together.

The disease consequences map onto specific nodes. Defective efferocytosis, often via proteolytic shedding of MerTK, drives necrotic core formation in atherosclerosis; C1q deficiency is the strongest single-gene risk factor for lupus; TREM2 variants impair microglial clearance of amyloid; and in tumours the CD47–SIRPα axis is reinforced by a TGF-β-rich environment. Identifying the responsible population — F4/80, CD68, CSF1R (CD115) — is the usual first step.

Key targets and matching reagents

Studying macrophage phagocytosis in vivo

Almost every node on this map is a surface receptor or a secreted bridging molecule, which makes the pathway unusually amenable to functional-grade blockade. Experiments fall into three groups.

1. Lifting the don't-eat-me brakes

Anti-CD47 (MIAP410) and anti-CD172a address the dominant axis directly, while anti-PD-L1 (10F.9G2), anti-PD-1 (29F.1A12), anti-CD24, anti-CD200 (OX-90), anti-MHC class I and anti-CD31 (390) cover the brakes that compensate when CD47 alone is blocked. Pair each with a matched IgG2a isotype control: these antibodies are themselves opsonins, so an isotype control is the only way to separate checkpoint blockade from Fc-mediated uptake of the coated target.

2. Manipulating the engulfment receptors

Anti-CD16/CD32 and anti-CD16-2 (9E9) dissect which activating FcγR carries a response; anti-CD11b, anti-CD11c and anti-CD18 (M18/2) block the complement-receptor arm, and anti-C5aR1 removes the priming signal that activates it. Anti-CD115 depletes the compartment outright; anti-F4/80 (CI:A3-1) and anti-CD68 identify what remains. Low endotoxin matters here: LPS alone primes the oxidative burst and upregulates FcγRIV.

3. Quantifying signalling and output

ELISA readouts turn a blocking experiment into a quantitative one: Syk, Btk, PLCγ2, PI3Kδ, PTEN and SHP-1 phospho-Tyr536 for the signalosome; Rab5, LAMP1, cathepsin D, LC3B and NOX2 for maturation and the burst; calreticulin, Gas6, MFG-E8 and TREM2 for recognition; and TNF, IL-1β, IL-10 and TGF-β for the verdict.

Featured products for this pathway

Frequently asked questions

For research use only. Not for use in diagnostic or therapeutic procedures.