Experimental prestorage filtration removes antibodies and decreases lipids in RBC supernatants mitigating TRALI in vivo

Abstract

Transfusion-related acute lung injury (TRALI) remains a significant cause of transfusion-related mortality with red cell transfusion. We hypothesize that prestorage filtration may reduce proinflammatory activity in the red blood cell (RBC) supernatant and prevent TRALI. Filters were manufactured for both small volumes and RBC units. Plasma containing antibodies to human lymphocyte antigen (HLA)-A2 or human neutrophil antigen (HNA)-3a was filtered, and immunoglobulins and specific HNA-3a and HLA-2a neutrophil (PMN) priming activity were measured. Antibodies to OX27 were added to plasma, and filtration was evaluated in a 2-event animal model of TRALI. RBC units from 31 donors known to have antibodies against HLA antigens and from 16 antibody-negative controls were filtered. Furthermore, 4 RBC units were drawn and underwent standard leukoreduction. Immunoglobulins, HLA antibodies, PMN priming activity, and the ability to induce TRALI in an animal model were measured. Small-volume filtration of plasma removed >96% of IgG, antibodies to HLA-A2 and HNA-3a, and their respective priming activity, as well as mitigating antibody-mediated in vivo TRALI. In RBC units, experimental filtration removed antibodies to HLA antigens and inhibited the accumulation of lipid priming activity and lipid-mediated TRALI. We conclude that filtration removes proinflammatory activity and the ability to induce TRALI from RBCs and may represent a TRALI mitigation step.

Figure 1

Filtration algorithm of HLA antibody-positive donors and antibody-negative controls. The algorithm depicts the experimental design of the donors employed for manufacture of RBC units that underwent experimental filtration using filters B, P, or M or standard LR (Pall BPF4 LR filter). Please note that there are 4 antibody-negative male controls per experimental filter and 4 antibody-negative donors, 2 male and 2 female for the BPF4 LR filter.

Figure 2

Experimental filtration mitigates the priming activity of human antibodies against HLA-A2 and HNA-3a. (A) Priming of the fMLF-activated, superoxide dismutase inhibitable respiratory burst (nmol O₂⁻/min) in PMNs from HLA-A2 homozygotes primed with FP, unmodified plasma with antibodies to HLA-A2, and filtered plasma with antibodies to HLA-A2. Experimental filtration mitigated the priming activity, which was not different from the FP-treated controls. (B) Priming of the fMLF-activated, superoxide dismutase inhibitable respiratory burst in HNA-3a⁺ PMNs primed with FP, unmodified plasma that contains antibodies to HNA-3a, and experimentally filtered plasma that contains antibodies to HNA-3a. Experimental filtration mitigated the HNA-3a priming activity. *P < .05 vs all groups, n = 5.

Figure 3

Experimental filtration mitigates ALI in plasma that contains major histocompatibility complex class I antibodies to OX27. The figure illustrates EBD leak, a measure of ALI, as a function of treatment group. Saline-treated rats (first event) did not demonstrate ALI irrespective of the second event: heat-treated (56°C for 30 minutes) FP, FP with OX27 (FP+OX27), filtered FP (F-FP), or filtered FP + OX27 (FP+OX27). In comparison, endotoxin (LPS) pretreatment did evidence ALI with plasma that contained OX27 (LPS/FP+OX27) as the second event, but no ALI was evident in LPS-treated rats infused with plasma (LPS/FP). Experimental filtration of the plasma with OX27 mitigated ALI (LPS/FP+OX27), and there was no ALI produced by filtered FP (LPS/FP). *P < .05 vs all groups, n = 5. BALF, bronchoalveolar lavage fluid.

Figure 4

Experimental filtration mitigates the accumulation of PMN priming activity and lipid priming activity in the supernatant of RBC units during routine storage. Priming of the fMLF-activated, superoxide dismutase inhibitable respiratory burst (nmol O₂⁻/min) is depicted as a function of experimental filtration groups and days of routine storage (A). The standard leukoreduced controls (standard LR filter) demonstrate the accumulation of PMN priming activity as a function of storage with significant activity reached by day 21 with a relative maximum on day 42 of storage (*P < .05 vs day 1 supernatant). In the 3 experimental filters (B, P, and M), there is only an increase in priming activity in the RBC supernatants on day 42 vs day 1 (*P < .05 vs day 1 supernatant). The priming activity from supernatants from RBC units that underwent standard LR was significant from the experimental filtration groups on both day 21 and day 42 (†P < .05 vs day 21 and day 42 supernatants from experimentally filtered RBC units; n ≥ 10 for each group). Experimental filtration inhibited the accumulation of lipid priming activity in the supernatant of RBC units during storage (B) depicted as the maximal rate of superoxide anion production to fMLF (nmol O₂⁻/min) as a function of treatment group. The lipid extracts from the supernatant of the RBCs that underwent standard LR or experimental filtration were not different from albumin-treated controls. However, extracts from the supernatant of RBCs that underwent standard LR demonstrated increased priming activity as compared with albumin-treated controls and extracts from day 0 and day 42 supernatants from experimentally filtered RBCs (*P < .05 vs albumin controls and lipid extracts from day 1 supernatants, †P < .05 vs day 42 extracts or the supernatants of experimentally filtered RBCs; n = 15). Lastly, the extracts from the supernatants of experimentally filtered RBC units showed greater priming activity vs the albumin-treated controls and the extracts from the day 0 experimentally filtered extracts (*P < .05 vs albumin controls and lipid extracts from day 1 supernatants).

Figure 5

Experimental filtration inhibits ALI induced by the supernatant from stored day 42 LR-RBCs. EBD leak is depicted as a function of treatment group. Normal saline (NS) treatment of rats as the first event did not result in ALI regardless of the second event: NS or the heat-treated supernatants from day 1 or day 42 plasma from units that underwent experimental filtration of standard LR. In LPS-treated rats (first event), NS did not cause ALI; however, the heat-treated supernatants from day 42 RBCs that underwent standard LR induced significant EBD leak/ALI. Conversely, none of the heat-treated supernatants from the experimentally filtered units on day 42 induced significant EBD leak/ALI (*P < .05 vs LPS/NS and NS/NS controls, n = 5 for each group).