Storage temperature determines platelet GPVI levels and function in mice and humans

Abstract

Platelets are currently stored at room temperature before transfusion to maximize circulation time. This approach has numerous downsides, including limited storage duration, bacterial growth risk, and increased costs. Cold storage could alleviate these problems. However, the functional consequences of cold exposure for platelets are poorly understood. In the present study, we compared the function of cold-stored platelets (CSP) with that of room temperature-stored platelets (RSP) in vitro, in vivo, and posttransfusion. CSP formed larger aggregates under in vitro shear while generating similar contractile forces compared with RSP. We found significantly reduced glycoprotein VI (GPVI) levels after cold exposure of 5 to 7 days. After transfusion into humans, CSP were mostly equivalent to RSP; however, their rate of aggregation in response to the GPVI agonist collagen was significantly lower. In a mouse model of platelet transfusion, we found a significantly lower response rate to the GPVI-dependent agonist convulxin and significantly lower GPVI levels on the surface of transfused platelets after cold storage. In summary, our data support an immediate but short-lived benefit of cold storage and highlight the need for thorough investigations of CSP. This trial was registered at www.clinicaltrials.gov as #NCT03787927.

© 2021 by The American Society of Hematology.

Figures

Graphical abstract

Figure 1.

Platelet aggregate force and area in response to shear flow. Reconstituted whole-blood samples with fresh (green triangle), RT-stored (red circles), or 4°C-stored (blue squares) apheresis platelets were perfused through a microfluidic device to measure the force and area of platelet aggregates formed under shear flow. (A) Each microfluidic channel contains multiple sets of block and post force sensors. While under flow, platelets attach and aggregate to form a plug-like structure (green) that encapsulates the block and post. Aggregated platelets are able to produce contractile forces (F) that pull the flexible post toward the rigid block. Force is calculated from displacement of the tip of the post (δ) using Hooke’s law: F = k δ, where k = 3π E d4/64 L3, and E is the modulus of elasticity, d is the diameter, and L is the length of the post. (B) Representative images of platelet aggregate area 15, 60, and 300 seconds after blood enters the channel. (C) Mean force of the platelet aggregates over time (n = 5; shaded regions represent standard error of the mean [SEM]). (D) Force of the platelet aggregates at 60 seconds (n = 5; shown as mean ± SEM). (E) Force of the platelet aggregates at 300 seconds (n = 5; shown as mean ± SEM). (F) Mean area of the platelet aggregates over time (n = 5; shaded regions represent SEM). (G) Area of the platelet aggregates at 60 seconds (n = 5; shown as mean ± SEM). (H) Area of the platelet aggregates at 300 seconds (n = 5; shown as mean ± SEM). *P = .0103 for fresh and RT (H) , **P = .0032 for 4°C and RT (G).

Figure 2.

Single-platelet contraction force and spread area. Apheresis platelets were washed and seeded onto flexible PDMS substrates that were printed with an array of black dots to measure traction forces and spread area of individual platelets. (A) Platelets were fixed, stained, and imaged to visualize F-actin (green), GPIb (purple), and the array of black dots (orange). The magnitude and direction of platelet traction forces (blue arrows) were calculated from the displacement of the dots. (B) Traction forces were measured for fresh (green), RT-stored (red), and 4°C-stored (blue) platelets that were seeded onto black dot substrates coated with VWF. Violin plots show data from a representative donor for whom 252 platelets were measured (fresh, n = 117; RT stored, n = 73; 4°C stored, n = 62). (C) Average traction forces per platelet were measured for 6 donors, and no statistically significant difference was observed between fresh (green triangles), RT-stored (red circles), and 4°C-stored (blue squares) platelets. (D) Average spread area of platelets on VWF-coated black dots was measured for each donor, and no significant difference was observed between the conditions. (E) Traction forces were measured for fresh (green), RT-stored (red), and 4°C-stored (blue) platelets on fibrinogen-coated black dots. Violin plots show data from a representative donor for whom 242 platelets were measured (fresh, n = 81; RT stored, n = 84; 4°C stored, n = 77). (F-G) Average traction forces (F) and average spread area (G) of platelets on fibrinogen-coated black dots were measured for 5 donors, and no significant difference was observed between conditions.

Figure 3.

Platelet storage temperature and response to agonists in PRP and washed platelets (WPs). We obtained human platelets by apheresis and used platelets either fresh (green triangles) or stored for 5 days at either 4°C (blue squares) or 22°C (RT; red circles). Aggregation was induced by stimulation with 5 μg/mL of collagen, 20 μM of ADP, or 0.5 mM of arachidonic acid (AA; shown as maximum aggregation; mean ± standard error of the mean [SEM]; n = 6-7). (A) PRP: collagen, ADP, and AA (left) and representative aggregation traces (right). (B) WPs: collagen, ADP, and AA (left) and representative aggregation traces (right). (C) GPVI levels on platelets determined by flow cytometry with fluorochrome-conjugated anti-GPVI antibody (n = 7; left), And β1 integrin levels on platelets determined by flow cytometry with fluorochrome-conjugated β1 antibody (n = 7; right). (D) Separate cohort of healthy volunteers, whose platelets were stored for 7 days at RT or 4°C under the same conditions as described for the original cohort. GPVI levels were determined using liquid chromatography–tandem mass spectrometry. Results are shown as fold change from baseline (fresh; n = 5). (E) PRP was diluted with separately stored plasma and stimulated with 100 ng/mL of convulxin. We stained with antibodies against activated integrin (PAC-1) and P-selectin (anti-CD62P; shown as change in mean fluorescence intensity [MFI] ± SEM from unstimulated; n = 5). *P < .5, **P < .01, ***P < .001. ns, not significant.

Figure 4.

Posttransfusion and in vivo platelet (PLT) function in healthy humans. (A) Time course of the healthy human crossover study, including baseline (BL) PLT test; PLT function 24 hours after loading dose (LD) with 325 mg of ASA and 600 mg of clopidogrel; and 1-, 4-, and 24-hour PLT function tests posttransfusion. (B) Absolute PLT counts at different time points of the study from individuals during RT-stored PLT transfusion round (red circles) and 4°C-stored PLT transfusion round (blue squares; n = 7-8; shown as mean ± standard error of the mean [SEM]). (C) Corrected count increments at posttransfusion time points (n = 7-8; shown as mean ± SEM). (D) PLTs were washed before assessment by light transmission aggregometry. Platelets were stimulated with 20 μg/mL of collagen, 20 mM of ADP, and 0.5 mM of arachidonic acid (n = 7-8; shown as mean ± SEM of maximum aggregation). (E) PLT reactivity tested by Verify NOW for ASA (aspirin reaction units [ARUs]; n = 7-8; shown as mean ± SEM). (F) Individual responses of participants to autologous transfusion, after LD (normalized to 0) and at 1, 4, and 24 hours posttransfusion (shown as percentage change from bleeding LD value; negative values indicate shortening of bleeding time [BT; ie, correction of prolonged BT], and positive values indicate prolongation of bleeding time [ie, worsening of prolonged BT]). Responses after transfusion of cold-stored autologous units (4°C stored; n = 8; left), and responses after transfusion of RT-stored autologous units (RT stored; n = 7; right). *P = .0110 (4-hour time point) and P = .0154 (24-hour time point) (B) and P = .015 (4-hour time point) and P = .0098 (24-hour time point) (C), **P = .0088 (D) and P = .0018 (E). CCI, corrected count increment; ns, not significant.

Figure 5.

Stored platelet (PLT) posttransfusion function in mice. (A) Outline of the mouse PLT transfusion model (details provided in “Methods”). (B) PLT in vivo survival after storage for 24 hours at either RT (red circles) or 4°C (blue squares; n = 9). (C) αIIbβ3 integrin activation in whole blood 4 and 24 hours after PLT transfusion with either 24-hour RT- (red circles) or 4°C-stored (blue squares) PLTs after stimulation with 100 nM of convulxin (n = 9; shown as mean fluorescence intensity [MFI] ± standard error of the mean [SEM] of Jon/A antibody binding). (D) GPVI expression 4 and 24 hours after transfusion of either 24-hour RT- (red circles) or 4°C-stored (blue squares) PLTs (n = 9; shown as MFI ± SEM of JAQ-1 with goat anti-rat immunoglobulin G–phycoerythrin secondary antibody). *P = .032 (4 hours) (C) and P = .038 (4 hours) (D), **P = .001 (4 hours: RT vs 4°C) (B), P = .012 (24 hours) (C), and P = .008 (24 hours) (D). CVX, convulxin; RBC, red blood cell; WT, wild type.