Antibodies against biotin-labeled red blood cells can shorten posttransfusion survival

Abstract

Background: In hematologic and transfusion medicine research, measurement of red blood cell (RBC) in vivo kinetics must be safe and accurate. Recent reports indicate use of biotin-labeled RBC (BioRBC) to determine red cell survival (RCS) offers substantial advantages over ⁵¹ Cr and other labeling methods. Occasional induction of BioRBC antibodies has been reported.

Study design and methods: To investigate the causes and consequences of BioRBC immunization, we reexposed three previously immunized adults to BioRBC and evaluated the safety, antibody emergence, and RCS of BioRBC.

Results: BioRBC re-exposure caused an anamnestic increase of plasma BioRBC antibodies at 5-7 days; all were subclass IgG₁ and neutralized by biotinylated albumin, thus indicating structural specificity for the biotin epitope. Concurrently, specific antibody binding to BioRBC was observed in each subject. As biotin label density increased, the proportion of BioRBC that bound increased antibody also increased; the latter was associated with proportional accelerated removal of BioRBC labeled at density 6 μg/mL. In contrast, only one of three subjects exhibited accelerated removal of BioRBC density 2 μg/mL. No adverse clinical or laboratory events were observed. Among three control subjects who did not develop BioRBC antibodies following initial BioRBC exposure, re-exposure induced neither antibody emergence nor accelerated BioRBC removal.

Discussion: We conclude re-exposure of immunized subjects to BioRBC can induce anamnestic antibody response that can cause an underestimation of RCS. To minimize chances of antibody induction and underestimation of RCS, we recommend an initial BioRBC exposure volume of ≤10 mL and label densities of ≤18 μg/mL.

Keywords: BioRBC antibodies; RBC posttransfusion kinetics; biotin; biotin-labeled RBC.

FIGURE 1

In the three nonimmunized control subjects, red cell survival (RCS) following biotin-labeled red blood cell (BioRBC) re-exposure was similar to initial BioRBC exposure. The BioRBC densities transfused in the initial and re-exposure-1 studies are indicated in the figure legend. Among the multiple BioRBC densities transfused (Table 1), only data for nonoverlapping, discreetly resolvable BioRBC densities (i.e., ≤54 μg/dL) are included. In response to BioRBC re-exposure-1, none developed BioRBC antibodies

FIGURE 2

For subject IMU#1, biotin-labeled red blood cell (BioRBC) re-exposure-1 using a mixture of three BioRBC densities elicited an early, high titer, anamnestic BioRBC antibody response that appeared concurrently with acceleration of BioRBC disappearance. (A). Time course of plasma BioRBC antibody titer. Prior to re-exposure-1, BioRBC plasma antibodies were undetectable. After re-exposure-1, antibodies were detected at 7 days; maximum titer (day 14) was about 100-fold greater than maximum titer for the initial exposure (day 140). (B). Red cell survival (RCS) for three BioRBC densities. Concurrent with anamnestic antibody appearance, RCS for all BioRBC densities (as indicated in the legend) decreased relative to initial BioRBC exposure. Likewise, RCS was decreased relative to BioRBC survival in controls and relative to the normal RCS in healthy adults. Rates of disappearance were density-dependent (54 > 18 >6 μg/mL) and hence were also dose-dependent. Disappearance of BioRBC-6 following re-exposure-1 was slower than those of 18 and 54; disappearance of BioRBC-6 was biphasic with approximately half removed by 3 week followed with the remaining BioRBC-6 disappearing at a rate similar to subject’s initial exposure to BioRBC-6, and similar to normal RCS

FIGURE 3

For subject IMU#1, re-exposure-2 using a single biotin-labeled red blood cell (BioRBC) density (BioRBC-2) elicited a brisk anamnestic antibody response but no acceleration of BioRBC-2 removal. (A). BioRBC re-exposure-2 elicited the early appearance of antibodies (day 7 and 5, respectively). Antibody titers exhibited similar time course and maximum titer to that of re-exposure-1. Baseline antibody titer for re-exposure-2 (3 y after re-exposure-1) was 8. (B). Red cell survival (RCS) for BioRBC-2 following re-exposure-2 was normal (and the same as for BioRBC-6 in this subject’s initial BioRBC exposure). BioRBC-6 RCS following re-exposure-1 is provided for comparison. These observations are consistent with our hypothesis that BioRBC removal is density dependent and exhibits a threshold for removal

FIGURE 4

For immunized subject IMU#1, re-exposure-3 to a mixture of biotin-labeled red blood cell (BioRBC)-2 and −6 did not illicit a titer rise above the already high baseline titer of 64 and yet did accelerate removal of the BioRBC-6 but not of BioRBC-2. (A). Antibody titers remained constant at 64 throughout the study. For comparison, the antibody titers for re-exposure-1 and −2 are also depicted. (B). Red cell survival (RCS) was normal for BioRBC-2. RCS for BioRBC-6 was shortened similar to BioRBC-6 in re-exposure-1

FIGURE 5

Substantially different spectrum of biotin-labeled red blood cell (BioRBC) antibody responses and red cell survival (RCS) demonstrated for immunized subjects IMU #2 and #3 compared to IMU#1. (A). BioRBC antibody titer response to re-exposure for IMU#1, IMU#2, and IMU #3 differed substantially. BioRBC re-exposures for IMU#2 and IMU#3 were conducted 8 and 22 years after initial BioRBC exposures, respectively. For both subjects, prestudy antibodies were undetectable using standard assay conditions. Post transfusion, maximum antibody titer for IMU#2 was three orders of magnitude greater than that for IMU#3. (B). RCS of BioRBC-2 and −6 for IMU#1, IMU#2, and IMU#3 demonstrated an inverse relationship with concurrent maximum antibody titer (i.e., the greater the titer, the shorter the RCS) and with BioRBC label density (BioRBC-6 removal was accelerated relative to BioRBC-2)

FIGURE 6

Antibody binding to biotin-labeled red blood cell (BioRBC) coincides with emergence of antibody in plasma and likely accelerates BioRBC-6 removal. Example of dual staining data for IMU#2 re-exposure-1 shows that increasing plasma antibody titer (Figure 5A) parallels an increase in the proportion of BioRBC-6 that bind increased antibody (as detected by human anti-IgG-FITC) and is temporally associated with acceleration of BioRBC-6 removal. We enumerated RBC in six regions: R1: Unlabeled RBC without IgG antibody; R2: Unlabeled RBC with antibody; R3: BioRBC-2 without antibody; R4: BioRBC-2 with antibody; R5: BioRBC-6 without antibody; R6: BioRBC-6 with antibody. Pre-BioRBC transfusion panel. Only 0.05% of the unlabeled RBC exhibited increased antibody binding (defined as R2/(R1 + R2)), and only a small number of RBC were detected in the BioRBC channels (i.e., R3, R4, R5, and R6 contained <500 of 10⁶ RBC counted). 20-min post-BioRBC transfusion panel. The BioRBC-2 and −6 were easily detected in R3 and R5. Few (≤0.22%) of the unlabeled RBC, BioRBC-2 (=R4/(R4 + R3)), or BioRBC-6 (=R6/(R6 + R5)) exhibited increased antibody binding. Nonspecific increased antibody binding (e.g., an effect of aging of RBC or nonspecific IgG response to BioRBC re-exposure) would have been detected as increases in R2, R4, and R6, which were not observed. Day 8 post-BioRBC transfusion panel. Coincident with anamnestic antibody response, increased antibody binding was observed for 28% of the BioRBC-6 and 1.4% of BioRBC-2. This provides evidence of dependence of antibody binding on label density. Antibody binding to unlabeled RBC remains at baseline values (0.06%) providing evidence of specificity for BioRBC antibody

FIGURE 7

Variable responses of the three immunized subjects when re-exposed to a mixture of autologous BioRBC-2 and −6: Effect on the percentage of RBC with increased antibody bound as detected by human anti-IgG-FITC and on red cell survival (RCS) of biotin-labeled red blood cell (BioRBC)-N. The enumeration areas are the same as in Figure 6. IMU#2 re-exposure-1 A1. Over the first 14 days posttransfusion of BioRBC, the percentage of BioRBC binding abnormal amounts of antibody increased approximately in parallel with antibody titer (Figure 5A). Concurrently, BioRBC rapidly disappeared, with BioRBC-6 being removed more rapidly than BioRBC-2. On day 5, the percentage of BioRBC-6 binding abnormally increased human anti-IgG-FITC was five-fold >BioRBC-2; this is consistent with label density-dependence of antibody binding and with BioRBC antibody binding as the primary cause of accelerated BioRBC removal. A2. By day 8, the numbers of BioRBC-6 remaining in circulation were too few (<1%) for accurate enumeration, thus rendering estimates of the percent with increased antibody BioRBC binding less precise; likewise for BioRBC-2 by day 14. IMU#3 re-exposure-1 B1. The percentage of BioRBC-N that bound abnormal amounts of antibody increased slowly in parallel with the rise in antibody titer. On both days 7 and 15, the percentage of BioRBC-6 with increased antibody binding was increased; although small relative to IMU#2, percentage of BioRBC-6 with increased antibody was still five-fold >BioRBC-2, thus providing evidence for label density-dependence. B2. Both BioRBC-6 and-2 survived normally through day15. However, examination of RCS over a longer period (Figure 5B) reveals that RCS of BioRBC-6 was shortened by day 29 and that BioRBC-2 had normal RCS. These observations are consistent with bound antibody as the primary determinant of BioRBC removal and for existence of a threshold for detectable shortening of RCS. IMU#1 re-exposure-3 C1. The plasma antibody titer for IMU#1 re-exposure-3 (Figure 5A) was already increased before transfusion of the BioRBC mixture and remained constant throughout. The percentage of BioRBC-2 and BioRBC-6 that bound abnormal amounts of antibody was already increased on Day 1 (2% and 37%, respectively). Maximum percent abnormal antibody binding for BioRBC-6 was about five-fold >BioRBC-2, consistent with label density-dependence. C2. BioRBC-2 had normal RCS, and BioRBC-6 exhibited definite shortening of RCS. Yet, a minority of the BioRBC-6 exhibited normal RCS, which is apparent with the longer observation period (Figure 5B). In summary, these observations are all consistent with our hypotheses that, for immunized subjects, antibody binding is the primary determinant of shortened RCS and that there is a threshold for this mechanism