Abstract

Transfusion related acute lung injury (TRALI) is a serious complication of blood administration. TRALI is neutrophil (PMN)-mediated and the pathogenesis is due to the infusion of antibodies directed against specific HLA (class I or II) or granulocyte antigens and may also be the result of two events: 1) a predisposing clinical condition of the patient resulting in PMN pulmonary sequestration followed by 2) the infusion of biologic response modifiers (BRMs), e.g. lipids from stored blood. These BRMs prime PMNs in vitro and cause PMN cytotoxicity in vivo resulting in endothelial damage, capillary leak and TRALI. We hypothesize that antibodies directed against specific PMN antigens cause rapid PMN priming and cytotoxicity in a two-event model, specific to the antigen they recognize. PMNs were isolated from healthy donors and 5b (HNA-3a) donors by standard techniques and were incubated for 3 min with 10% fresh plasma (FP) from healthy donors, plasma from three donors who had antibodies against HNA-3a and were implicated in TRALI, or 10% immune complexes (ICs) made from sera (+ control). Selected wells had Fab′2 fragments against the Fc receptors (CD16, CD32 & CD64). The maximal rate of O2 production was measured as cytochrome c reduction at 550 nm. Priming is defined as augmentation of the fMLP-activated respiratory burst (Table 1). We used an in vitro model of TRALI (Wyman AJP Cell 283: C1592, 2002) in which Human plumonary microvascular endothelial cells (HMVECs) were grown to 90% confluence and incubated with buffer or endotoxin (LPS) [200 ng/ml] for 6 hours. HNA-3a+ PMNs were added ± Fab′2 fragments, allowed to settle, and incubated for 30 min with FP or HNA-3a+ plasma. The plates were forcibly decanted and the viable HMVECs per mm2 were counted (Table 2). HNA-3a+ plasma significantly primed the PMN oxidase as compared to FP-treated controls (Table 1). Pre-treatment of PMNs with Fab′2 inhibited IC priming but did not affect other groups, and HNA-3a+ plasma did not prime HNA-3a PMNs (data not shown). HNA-3a+ plasma ± LPS without PMNs did not affect HMVEC viability (data not shown), and HMVECs activated with LPS caused widespread PMN adherence. HNA-3a+ plasma plus HNA-3a+ PMNs caused destruction of LPS-activated HMVECs that was partially inhibited by Fab′2 fragments (Table 2). We conclude that HNA-3a+ plasma rapidly, effectively, and specifically primes HNA3+ PMNs. HNA-3a+ plasma can also serve as the second event in a two-event model of PMN-mediated HMVEC damage. Thus, antibodies directed against specific granulocyte antigens and lipids from stored blood may cause TRALI through a common final pathway of PMN activation.

HNA-3a+ (5b+) plasma priming of HNA-3a+ PMNs

Pre-tx/GroupControlICsFFPDonor 1Donor 2Donor 3
Units=nmol O2/min; †=P<0.05 versus FFP controls; *=p<0.05 versus ICs. 
Buffer 2.7±0.5 8.8±0.7† 2.5±0.4 5.7±1.8† 6.5±† 5.0±1.0† 
Fab′2 2.1±0.8 3.2±0.4* 3.6±0.9 3.8±0.7 4.6±1.5 4.0±1.0 
Pre-tx/GroupControlICsFFPDonor 1Donor 2Donor 3
Units=nmol O2/min; †=P<0.05 versus FFP controls; *=p<0.05 versus ICs. 
Buffer 2.7±0.5 8.8±0.7† 2.5±0.4 5.7±1.8† 6.5±† 5.0±1.0† 
Fab′2 2.1±0.8 3.2±0.4* 3.6±0.9 3.8±0.7 4.6±1.5 4.0±1.0 

HNA-3a+ (5b+) plasma causes PMN cytotoxicity

Pre-tx/GroupControlLPS/PMNsLPS/PMNs/FFPLPS/PMNs/5b plasma
Units=viable HMVECs/mm2; †=p<0.05 versus Controls and LPS/PMNs/FFP; *=p<0.05 versus LPS/PMN/5b+ plasma. 
Buffer 245±1 234±8 229±9 151±7† 
Fab′2 no data no data 240±9 190±* 
Pre-tx/GroupControlLPS/PMNsLPS/PMNs/FFPLPS/PMNs/5b plasma
Units=viable HMVECs/mm2; †=p<0.05 versus Controls and LPS/PMNs/FFP; *=p<0.05 versus LPS/PMN/5b+ plasma. 
Buffer 245±1 234±8 229±9 151±7† 
Fab′2 no data no data 240±9 190±* 

Author notes

Corresponding author