AUTHORS
Nikhil Ranjan, PhD, Michael Arthur Cole, MD, DPhil2, Gloria Gerber, MD3, Daniel Flores-Guerrero4, Shruti Chaturvedi Jr, MBBS, MS5, Robert Brodsky, MD6
AFFILIATIONS
1 Johns Hopkins University, Baltimore, Maryland, United States
2 Johns Hopkins University, Baltimore, Maryland, United States
3 Johns Hopkins University, Baltimore, Maryland, United States
4 Johns Hopkins University, Baltimore, Maryland, United States
5 Johns Hopkins University, Baltimore, Maryland, United States
6 Johns Hopkins University, Baltimore, Maryland, United States
ABSTRACT
Introduction
Paroxysmal nocturnal hemoglobinuria (PNH) erythrocytes lack the GPI-anchored regulators CD55 and CD59 and are destroyed by the membrane attack complex (MAC). Complement inhibitors have transformed survival but do not normalize red cell lifespan, and the intracellular events that follow MAC insertion are poorly defined. Chen et al. (Cell 2025) reported that complement-exposed erythrocytes are lysed by an inflammatory programmed cell death ("spectosis"), in which a truncated NLRP3 (miniNLRP3) nucleates ASC and activates caspase-8 to cleave beta-spectrin, implying that inflammasome or caspase inhibition could mitigate hemolysis. This model has not been tested in primary erythrocytes from a PNH cohort under buffer conditions optimized for alternative pathway (AP) amplification.
Methods
Erythrocytes from 22 PNH patients (median Type III RBC clone 43.2%, range 0.7-99.7%) were washed and resuspended at 2% hematocrit. Patients transfused within 4 months were excluded from functional assays. Lysis was induced with 20% ABO-matched normal human serum (NHS) in acidified GVB0-3mM MgEGTA, pH 6.4, for 2 h at 37C, and free hemoglobin read at 541 nm. Saline, Mg-only, and MgEGTA buffers were compared directly. Cells or serum were pretreated 30 min with eculizumab (338 nM), Factor D inhibitor (FDi, 1 µM), sutimlimab (41 nM), Z-VAD-FMK (10 µM), Z-IETD-FMK (10 µM), MCC950 (10 µM), calpeptin (50 µM), PD150606 (50 µM), or epoxomicin (1 µM). Potassium efflux was tracked with IPG-4 AM (3.5 µM) in TF-1 and TF-1 PIGA-null cells, with aerolysin as a GPI-dependent pore control. For immunoblot, whole packed cells were lysed directly in LDS buffer, loaded volumetrically (1.3 uL packed RBC/lane), and normalized to total protein stain. NLRP3 was probed with 2 antibodies against separate domains (PYD, NACHT) across 105 to 107 cell equivalents of ghosts and whole packed RBCs; THP-1 lysate served as positive control.
Results
Active serum induced hemolysis proportional to clone burden, 73% in patients with Type III clones >50% (n=11) compared to 19% in clones <50% (n=11). Eculizumab and FDi rescued both cohorts (66-85% reduction, p<0.0001). In contrast, Caspase-8 inhibition (Z-IETD-FMK), pan-caspase inhibition (Z-VAD-FMK) and NLRP3 inhibition (MCC950) failed to protect.
MAC insertion drove rapid potassium (K+) efflux in TF-1 PIGA-null cells that was abrogated by complement blockade; caspase and NLRP3 inhibitors only blunted the early phase. TF-1 wild-type cells were protected by endogenous CD59, and aerolysin produced the reciprocal pattern.
NLRP3 was undetectable in ghosts and in whole packed RBCs with both antibodies, including the 70-90 kDa window predicted for an LRR-truncated variant, while THP-1 controls and total protein stain confirmed adequate loading. A published deep RBC proteome (>5,200 proteins) also returned no NLRP3 peptides but identified ASC and pro-caspase-8 at 5-7 unique peptides each, so erythrocytes retain the adaptor and effector but not the nucleating sensor. The Cryo-2 clone bound a prominent ~260 kDa species in ghost fractions, co-migrating with spectrin.
Active NHS produced beta-spectrin cleavage in patients with clones >90% (1.6-fold over heat-inactivated serum, p<0.0001); no cleavage was seen in clones <5%. Cleavage was prevented by complement blockade and by 2 structurally distinct calpain inhibitors, calpeptin and PD150606, but not by Z-VAD-FMK, epoxomicin, or mechanical lysis. Calpain inhibitors did not prevent hemolysis, uncoupling osmotic death from cytoskeletal degradation. Alpha-spectrin resisted complement but was cleaved by calcium ionophore A23187 in a calpeptin and PD150606 sensitive manner, indicating localized calcium entry through MAC pores is sufficient for beta-spectrin proteolysis but below the threshold for alpha-spectrin.
Conclusions
In primary PNH erythrocytes, complement-mediated destruction proceeds through a MAC-calcium-calpain axis rather than an NLRP3-caspase-8 programmed death cascade. We detect no NLRP3 protein in mature erythrocytes and no protection from inflammasome or pan-caspase blockade. Calpain-dependent beta-spectrin cleavage is dissociable from lysis, raising the possibility that sublytic cytoskeletal injury in cells surviving complement inhibition contributes to reduced deformability and extravascular clearance. These data indicate that targeting inflammasome or caspase signaling is unlikely to prevent complement-mediated erythrocyte destruction in PNH.
DISCLOSURES
Nikhil Ranjan, PhD: No relevant conflicts of interest to declare.