Magrolimab (Hu5F9-G4; Gilead Sciences) is an investigational humanised IgG4 monoclonal antibody (Gilead stopped the Phase 3 ENHANCE-3 AML study for futility and the FDA placed a clinical hold on magrolimab AML/MDS studies in February 2024) that binds the extracellular domain of CD47 and blocks its interaction with SIRPα (Signal Regulatory Protein alpha; CD172a) on macrophages, monocytes, neutrophils, and dendritic cells. By abolishing the CD47-SIRPα 'don't eat me' signal, magrolimab removes the myeloid phagocytosis brake and permits tumour-associated macrophages (TAMs) and neutrophils to phagocytose and clear tumour cells — a mechanism entirely distinct from T-cell checkpoint blockade (PD-1/LAG-3) or metabolic immunosuppression axes (adenosine, IDO1, succinate direct suppression). **SDH-specific mechanistic rationale — constitutive HIF-driven CD47 overexpression:** SDH loss → succinate accumulation → competitive inhibition of PHD2 and PHD3 prolyl hydroxylases → VHL-independent HIF-1α/2α stabilisation → constitutive pseudohypoxia. Samanta et al. (PNAS 2018, PMID 29367423, DOI 10.1073/pnas.1718197115) reported, in chemotherapy-enriched triple-negative breast cancer cells (a non-SDH model), that HIF-1α simultaneously co-induces CD47, CD73, and PD-L1 on tumour cells via canonical hypoxia-response elements (HREs) in each gene's promoter. In hypoxic cancer cells: (1) CD47 mRNA and protein expression were significantly upregulated; (2) this upregulation was HIF-1α-dependent (attenuated by HIF-1α silencing); (3) CD47-mediated phagocytosis resistance was dramatically reduced by anti-CD47 antibody treatment. In SDH-deficient tumours, where succinate-driven PHD inhibition creates constitutive, oxygen-independent HIF-1α activity, CD47 is predicted to be constitutively overexpressed via the same HRE mechanism (untested in SDH-deficient systems). HIF-1α therefore simultaneously drives three immunosuppression arms in SDH-deficient cells: the PD-L1/PD-1 T-cell checkpoint (existing engine entry), the CD73/adenosine T-cell suppression axis (oleclumab entry), and the CD47/SIRPα macrophage phagocytosis-evasion axis (this entry — the first myeloid axis in the engine). **CD47-SIRPα biochemical mechanism:** SIRPα carries cytoplasmic immunoreceptor tyrosine-based inhibitory motifs (ITIMs). CD47-SIRPα engagement triggers ITIM tyrosine phosphorylation, recruiting the Src homology 2 domain-containing phosphatases SHP-1 (PTPN6) and SHP-2 (PTPN11). These phosphatases dephosphorylate and inactivate myosin IIA — a motor protein essential for the cytoskeletal reorganisation, membrane protrusion, and contractile forces required to form a phagocytic cup and complete engulfment. Net effect: TAMs and neutrophils cannot engulf CD47-high tumour cells. Magrolimab occupies the CD47-SIRPα binding interface, blocking ITIM signalling and restoring the macrophage's intrinsic phagocytic capacity. **Mechanistic distinction from existing engine immune entries:** - **PD-1/PD-L1 axis** (existing entries): adaptive T-cell immune checkpoint; acts on infiltrating CD8⁺ T cells via inhibitory PD-1 ligation. CD47/SIRPα is an innate immune macrophage checkpoint; targets a completely different immune cell type (myeloid) and distinct molecular machinery. - **LAG-3 (relatlimab)**: adaptive T-cell exhaustion co-inhibitory checkpoint; CD47/SIRPα targets innate phagocytosis and is mechanistically orthogonal. - **Oleclumab (anti-CD73) / etrumadenant (A2AR antagonist)**: target the adenosine-mediated T-cell suppression axis; CD47/SIRPα is completely adenosine-independent. - **cGAS-STING**: innate immune sensing of cytosolic dsDNA → type I interferon production; distinct from myeloid phagocytosis checkpointing. - **IDO1/epacadostat**: tryptophan-kynurenine T-cell suppression; different axis. Magrolimab is the only engine entry targeting the myeloid (macrophage/neutrophil) phagocytosis-checkpoint axis. **Combination rationale:** Magrolimab is hypothesised to combine favourably with tumour-opsonising antibodies (which provide Fcγ receptor-activating 'eat me' signals) and with PD-1/PD-L1 inhibitors (simultaneous myeloid phagocytosis and T-cell activation). In SDH-deficient tumours, HIF-1α co-induction of CD47 and PD-L1 (both from PMID 29367423) makes simultaneous CD47 + PD-1/PD-L1 blockade a particularly logical combination, targeting two co-regulated HIF-driven immunosuppression mechanisms in the same tumour cells. **Clinical data:** - NCT02216409 (Phase 1b; Stanford University / Forty Seven Inc.; solid tumours including colorectal, ovarian, NSCLC, bladder, breast; n=62 phase 1b expansion; COMPLETED): First-in-human Phase 1b safety, PK, and preliminary efficacy of magrolimab in solid tumours. Established on-target anaemia (from phagocytosis of red blood cells expressing CD47) as the primary toxicity — managed with a 1 mg/kg priming dose to clear senescent RBCs before therapeutic dosing — and preliminary pharmacodynamic target engagement. **Key limitation:** No published data test magrolimab or any anti-CD47 agent specifically in SDH-deficient GIST, PPGL, RCC, or pituitary adenoma models. CD47 protein expression in SDH-deficient tumour tissue has not been quantified by IHC or confirmed to be HIF-dependently elevated. The mechanistic rationale rests on the HIF-1α → CD47 HRE axis (PMID 29367423) established in non-SDH hypoxic cancer cells and the constitutive pseudohypoxia of SDH-deficient tumours. On-target anaemia risk requires a priming dose strategy to be clinically feasible. Required next steps: CD47 IHC on SDH-deficient GIST/PPGL TMA panels; CD47 qPCR in SDH-null vs. SDH-intact isogenic cell lines; phagocytosis co-culture assay with magrolimab + TAM-like macrophages and SDH-deficient cancer cells. Evidence_score 22 (theoretical): strong HIF→CD47 mechanistic chain established by PMID 29367423 in non-SDH hypoxic cancer cells; constitutive pseudohypoxia in SDH-deficient tumours makes constitutive CD47 overexpression biologically plausible; Phase 1b clinical pharmacology data (NCT02216409) establish human tolerability; score held at 22 due to complete absence of SDH-specific preclinical or clinical data and unconfirmed CD47 IHC in SDH-deficient tumour tissue.