Rapid clearance of storage-induced microerythrocytes alters transfusion recovery

Abstract

Permanent availability of red blood cells (RBCs) for transfusion depends on refrigerated storage, during which morphologically altered RBCs accumulate. Among these, a subpopulation of small RBCs, comprising type III echinocytes, spheroechinocytes, and spherocytes and defined as storage-induced microerythrocytes (SMEs), could be rapidly cleared from circulation posttransfusion. We quantified the proportion of SMEs in RBC concentrates from healthy human volunteers and assessed correlation with transfusion recovery, investigated the fate of SMEs upon perfusion through human spleen ex vivo, and explored where and how SMEs are cleared in a mouse model of blood storage and transfusion. In healthy human volunteers, high proportion of SMEs in long-stored RBC concentrates correlated with poor transfusion recovery. When perfused through human spleen, 15% and 61% of long-stored RBCs and SMEs were cleared in 70 minutes, respectively. High initial proportion of SMEs also correlated with high retention of RBCs by perfused human spleen. In the mouse model, SMEs accumulated during storage. Transfusion of long-stored RBCs resulted in reduced posttransfusion recovery, mostly due to SME clearance. After transfusion in mice, long-stored RBCs accumulated predominantly in spleen and were ingested mainly by splenic and hepatic macrophages. In macrophage-depleted mice, splenic accumulation and SME clearance were delayed, and transfusion recovery was improved. In healthy hosts, SMEs were cleared predominantly by macrophages in spleen and liver. When this well-demarcated subpopulation of altered RBCs was abundant in RBC concentrates, transfusion recovery was diminished. SME quantification has the potential to improve blood product quality assessment. This trial was registered at www.clinicaltrials.gov as #NCT02889133.

Graphical abstract

Figure 1.

Accumulation of morphologically altered RBCs during storage. (A) Panoramic (×2000) (i) and detailed (×5000) (ii) views of short-stored RBC sample (day 12 of storage) containing a majority of discocytes (D) and type I (EI) or II echinocytes (EII), and panoramic (×2000) (iii) and detailed (×5000) (iv) views of long-stored RBC sample (day 41 of storage) containing abundant EIII, spheroechinocytes (SE), and spherocytes (S). (B) Representative scanning electron images of RBC shapes observed during storage in SAGM: D, EI, EII, EIII, SE, and S; numerically zoomed regions from 2000× acquisitions. Red square highlights morphologically altered RBCs defined as SMEs. Scale bars = 2 μM.

Figure 2.

Proportion of SMEs at the end of storage correlates with 24-hour posttransfusion recovery in healthy human volunteers. (A) Quantification of SMEs upon storage of RBC concentrates in SAGM solution (n = 24) between days 3 and 42 (mean value in solid black line). Low (blue lines) and high proportions of SMEs (red lines) defined by proportions of SME < −1 SD (11%) and > +1 SD (38%) at the end of storage, respectively. (B) Representative normalized frequency plot for RBC concentrate at the end of storage in AS-3 showing a well-demarcated subpopulation of SMEs. Subpopulation of SMEs contains spherocytes, spheroechinocytes, and type III echinocytes (i), whereas normal-sized RBCs (ii) contain discocytes and type I and II echinocytes. (C) Correlation between 24-hour posttransfusion recovery and proportions of SMEs quantified by imaging flow cytometry at the end of storage (n = 31; P = .02; Spearman r = −0.42; r² = 0.24).

Figure 3.

SMEs are rapidly cleared when perfused through human spleen ex vivo. (A) Kinetics (means ± standard errors of the mean) of the normalized concentration in the perfusate of human spleen ex vivo (n = 7) of 14 RBC concentrates stored for 35 to 42 days and rejuvenated (RW; dashed line) or not (NT; solid line). (B) Representative normalized frequency plot of the projected surface area of RBCs stored for 42 days before (solid red histogram) and after (dotted line) rejuvenation. Dashed vertical line defines the gating cutoff for SMEs. (C) Representative frequency plot of projected surface area of stored RBCs (37 days) before (red histogram) and at different time points after perfusion through human spleen ex vivo (0 minutes, solid blue line; 2 minutes, solid green line; 20 minutes, dotted black line; and 40 minutes, dashed orange line). (D) Proportion of SMEs at the beginning (T0) and mean proportion of SMEs at all observations between 40 and 70 minutes (T40-70min) of perfusion through human spleen ex vivo (n = 6; red dashed line represents mean). (E) Correlation between mean retention rate in human spleen perfused ex vivo and proportion of SMEs in the RBC concentrate before transfusion (n = 28; P = .03; Spearman r = 0.5; r² = 0.2). *P < .05, **P < .01 by Sidak multiple comparisons test comparing, at each time point, the persistence in circulation of rejuvenated vs nontreated RBCs (A) or by paired Student t test comparing proportion of SMEs at T0 vs T40-70min (D).

Figure 4.

Identification and quantification of a subpopulation of SMEs in a mouse model. (A) Posttransfusion recovery of long-stored RBCs (14 days of storage; lavender line; n = 7) is decreased at 2 and 24 hours after transfusion to control recipients compared with short-stored RBCs (1 day of storage; red line; n = 7). (B) Normalized frequency plots of projected surface area for fresh (purple line; n = 11) and short-stored RBCs (red line; n = 9) show similar patterns. (C) Long-stored RBCs (lavender line; n = 9) have a reduced projected surface area compared with short-stored RBCs (red line). (D) Quantification of projected surface area of front views of focused RBCs obtained by imaging flow cytometry shows a significant decrease in long-stored RBCs. (E) Representative analysis of brightfield images of long-stored RBCs shows that morphologically altered RBCs (dashed red line) have a reduced projected surface area compared with RBCs with normal morphology (blue line). Morphologically altered RBCs defined as type III echinocytes, spheroechinocytes, spherostomatocytes, and spherocytes, whereas normal RBCs comprised discocytes and type I and II echinocytes. (F) SMEs in short- and long-stored RBCs accumulate during storage. Data are presented as means ± standard errors of the mean. ***P < .001, ****P < .0001 by Sidak multiple comparisons test comparing, at each time point, recovery of short- vs long-stored RBCs (A), by Tukey multiple comparisons test comparing projected surface area for each condition (D), or by Student t test comparing proportion of SMEs in short-stored vs long-stored RBC (F). ns, not significant.

Figure 5.

In the mouse model, SMEs are rapidly cleared, accumulate in the spleen, and are processed predominantly by macrophages. (A) Representative normalized frequency plot of projected surface area for long-stored mouse RBCs, as observed at 5 minutes (green line), 2 hours (yellow line), and 24 hours (red line) after transfusion to a syngeneic C57Bl/6 mouse. Control fresh RBCs from a nontransfused mouse (blue) are shown as reference. Dashed black vertical line defines the gating of SMEs. (B) Declining proportion of SMEs in circulation after transfusion (n = 10 mice per group). (C) Variable persistence in circulation of long-stored RBCs (lavender line) that contain the 2 complementary subpopulations of SMEs (black dotted line) and morphologically normal RBCs (light blue solid line), computed by combining flow cytometric and Imagestream data (n = 10 mice per group). (D) Proportion of long-stored RBCs that were cleared 5 to 240 minutes posttransfusion (n = 8 mice per time point). (E) EF 5 to 240 minutes posttransfusion (ratio of transfused CFSE⁺ RBCs in sliced organ/CFSE⁺ RBCs in venous blood; n = 4 mice per time point) in spleen (red line), liver (orange line), and bone marrow (blue line). (F) Posttransfusion erythrophagocytosis of RBCs in spleen (i), liver (ii), and bone marrow (iii), estimated by the increase in CFSE median fluorescence intensity (MFI) of macrophages (red lines), monocytes (blue lines), inflammatory monocytes (purple lines), and granulocytes (orange lines), compared with control nontransfused mice (n = 3 mice per time point). Data are presented as means ± standard errors of the mean. *P < .05, ****P < .0001 by Kruskal-Wallis test compared with fresh RBC condition (B) or by Sidak multiple comparisons test comparing, at each time point, recovery of SME subpopulation vs normal subpopulation (C).

Figure 6.

Clearance kinetics of SMEs in clodronate-treated mice. (A) Representative normalized frequency plots of projected surface area for long-stored mouse RBCs, as observed 5 minutes (green line), 2 hours (yellow line), and 24 hours (red line) after transfusion of a clodronate-treated mouse. Control fresh RBCs from a nontransfused mouse (blue) are shown as reference. Dashed black vertical line defines gating of SME. (B) Delayed clearance of SMEs in circulation after transfusion in clodronate-treated mice (n = 10 per group). (C) Posttransfusion recovery of long-stored RBCs is increased at 2 and 24 hours after transfusion to clodronate-treated recipients (dashed line; n = 10) compared with controls (solid line; n = 10). (D) Variable persistence in circulation of long-stored RBCs (lavender dashed line) that contain the 2 complementary subpopulations of SMEs (black dotted line) and morphologically normal RBCs (light blue solid line), computed by combining flow cytometric and Imagestream data after transfusion in clodronate-treated mice (n = 10 per group). (E) Decreased proportion of long-stored RBCs that were cleared 60 to 240 minutes posttransfusion in clodronate-treated recipients (dashed line; n = 6 mice per time point) compared with controls (solid line; n = 8 mice per time point). (F) EF (ratio of transfused CFSE⁺ RBCs in sliced organ/CFSE⁺ RBCs in venous blood) 5 to 240 minutes posttransfusion in spleen (red line), liver (orange line), and bone marrow (blue line) after transfusion in clodronate-treated mice (n = 3 mice per time point). (G) Posttransfusion erythrophagocytosis of RBCs in spleen (i) and liver (ii), estimated by the increase in MFI of CFSE in monocytes (blue lines), inflammatory monocytes (purple lines), and granulocytes (orange lines) in clodronate-treated recipients compared with control nontransfused mice (n = 3 mice per time point). Data are presented as means ± standard errors of the mean. **P < .01, ***P < .001, ****P < .0001 by Kruskal-Wallis test compared with fresh RBC condition (B), by Sidak multiple comparisons test comparing, at each time point, recovery (clearance) in clodronate-treated vs control recipients (C,E), and by Sidak multiple comparisons test comparing, at each time point, recovery of SME subpopulation vs normal subpopulation (D).

Figure 7.

Clearance kinetics of SMEs in splenectomized mice. (A) Typical normalized frequency plots of projected surface area for long-stored mouse RBCs, as observed 5 minutes (green line), 2 hours (yellow line), and 24 hours (red line) after transfusion to a splenectomized mouse. Control fresh RBCs from a nontransfused mouse (blue) are shown as reference. Dashed black vertical line defines gating of SME. (B) Delayed clearance of SMEs in circulation after transfusion in splenectomized mice (n = 10 per group). (C) Posttransfusion recovery of long-stored RBCs is increased at 24 hours after transfusion to splenectomized recipients (dashed line; n = 10) compared with controls (solid line; n = 10). (D) Variable persistence in circulation of long-stored RBCs (lavender dashed line) that contain the 2 complementary subpopulations of SMEs (black dotted line) and morphologically normal RBCs (light blue solid line), computed by combining flow cytometric and Imagestream data, after transfusion in splenectomized mice (n = 10 per group). (E) Decreased proportion of long-stored RBCs that were cleared at 240 minutes posttransfusion in splenectomized recipients (dashed line; n = 6 mice per time point) compared with controls (solid line; n = 8 mice per time point). (F) EF (ratio of transfused CFSE⁺ RBCs in sliced organ/CFSE⁺ RBCs in venous blood) 5 to 240 minutes posttransfusion in liver (orange line) and bone marrow (blue line) after transfusion in splenectomized mice (n = 3 mice per time point). (G) Posttransfusion erythrophagocytosis of RBCs in liver (i) and bone marrow (ii), estimated by the increase in MFI of CFSE in macrophages (red lines), monocytes (blue lines), inflammatory monocytes (purple lines), and granulocytes (orange lines) in splenectomized recipients compared with control nontransfused mice (n = 3 mice per time point). Data are presented as means ± standard errors of the mean. *P < .05, **P < .01, ***P < .001, ****P < .0001 by Kruskal-Wallis test compared with fresh RBC condition (B), by Sidak multiple comparisons test comparing, at each time point, recovery (clearance) in splenectomized vs control recipients (C,E), and by Sidak multiple comparisons test comparing, at each time point, recovery of SME subpopulation vs normal subpopulation (D).