Release of biologically active CD154 during collection and storage of platelet concentrates prepared for transfusion


Neil Blumberg, Professor of Pathology and Laboratory Medicine, Director, Transfusion Medicine/Blood Bank, University of Rochester Medical Center, Box 608, 601 Elmwood Avenue, Rochester, NY 14642, USA.
Tel.: +1 585 275 3189; fax: +1 585 273 3002; e-mail:


Summary.  Background:  Millions of platelet transfusions are given each year. Transfusion reactions occur in as many as 30% of patients receiving unmodified platelet transfusions. The cause of some transfusion reactions remains unclear. The current paradigm suggests that platelet concentrates (PC) contain proinflammatory mediators that are released by white blood cells during collection, processing and storage. CD154 (CD40 ligand, CD40L) is a potent inflammatory mediator, normally sequestered inside the resting platelet, that is known to translocate to the platelet membrane and be shed into plasma in response to agonist activation. We hypothesized that platelet-soluble CD154 (sCD154) is ‘spontaneously’ released by transfused platelets and plays a major role in transfusion reactions.

Objectives:  To determine the time course and biological properties of CD154 translocation and release during collection and storage of platelets for transfusion.

Methods:  We measured surface and sCD154 in platelets prepared by the platelet-rich plasma method or apheresis by fluorescence-activated cell sorting and enzyme-linked immunosorbent assay, respectively. The specific biological activity of platelet sCD154 was assayed by stimulation of the CD154/CD40 pathway in known CD40-positive cells with PC-derived supernatants.

Results and conclusions:  We demonstrate that PCs prepared for transfusion have high levels of membrane-bound CD154 and sCD154, with maximum levels being seen 72 h after platelet collection. Importantly, we show that platelet-derived sCD154 potently stimulates CD40-positive cells. We propose that platelet-derived CD154 is a key ‘cytokine’ responsible for adverse reactions associated with platelet transfusions. Improved methods of platelet collection and/or storage, which limit CD154 expression, could reduce the risks of transfusion reaction.


CD154 (CD40 ligand, CD40L), a member of the tumor necrosis family, has been extensively studied for its role in adaptive immunity. Several important immunologic phenomena are attributed to CD154 interaction with its receptor, CD40, including B-cell immunoglobulin class switching, and maturation of dendritic cells [1,2]. Furthermore, the CD154–CD40 pathway is not limited to cells of the immune system. CD40 expression has been documented on non-hematopoetic cells, such as fibroblasts and keratinocytes [3,4], and recently identified on the surface of platelets [5]. Platelet agonist stimulation translocates endogenous CD154 to the cell membrane, where it interacts with CD40, causing cleavage of CD154 to generate the soluble biologically active product, soluble CD154 (sCD154) [6].

The significant discovery that platelets express CD154 augments their role as key players in inflammation [6]. For example, platelets induce expression of tissue factor and reduce expression of thrombomodulin in CD40-positive vascular cells in a CD154-dependent manner [7]. Platelets activated via CD154 signal vascular endothelial cell induction of chemokines [interleukin-8 (IL-8) and monocyte chemoattractant protein-1 (MCP-1)], adhesion molecules (E-selectin, vascular cell adhesion molecule type 1, and intercellular adhesion molecule type 1), cyclooxygenase-2 (COX-2) gene expression, and the expression of matrix metalloproteinases.

Platelet CD154 also has prothrombotic activity by virtue of its integrin-recognition sequence (KGD), which is important for binding to the major platelet integrin, αIIbβ3 [8], the primary receptor for fibrinogen on platelets. Platelet–platelet interaction through CD154–αIIbβ3 promotes stable platelet aggregation, and is necessary for stability of arterial thrombi [8]. Platelet-expressed CD154 has been observed in vivo in thrombi and atherosclerotic plaques, and disruption of CD40 diminishes the formation and progression of atherosclerotic plaques [6,8,9]. Platelets are the predominant source of circulating sCD154, and elevated levels are reported for a number of cardiovascular conditions and are associated with increased risk of future cardiovascular events in otherwise healthy women [10–12].

Platelets play a vital role in maintaining vascular integrity. Millions of platelets are transfused each year to increase platelet counts and reduce bleeding times, and to patients rendered thrombocytopenic by hematologic disease or cancer. Platelet transfusions may be accompanied by febrile non-hemolytic transfusion reactions (FNHTRs), characterized by fever, chills, and rigors [13]. Occasionally, life-threatening pulmonary failure occurs as a consequence of transfusion-associated acute lung injury syndrome (TRALI). Although severe reactions are infrequent, the reported incidence of mild FNHTRs ranges from 4% to 30% [13,14].

It was suggested that the pathogenesis of FNHTRs results from cytokines secreted by contaminating leukocytes during storage [15]. Prestorage leukoreduction of platelets diminishes cytokine (e.g. IL-1β, IL-6, and IL-8) accumulation in platelet concentrates (PCs), which correlates with a reduced adverse reaction rate [15,16]. The age of the platelet transfusion product is also associated with FNHTR incidence, with highest incidences of reactions being from non-leukoreduced platelets stored for 5 days [13,17]. Notably, prestorage leukoreduction of platelet products reduces but does not eliminate FNHTRs [16,18], implying that some mediators of FNHTRs are platelet-derived [19,20].

Materials and methods

Platelet preparation and activation

Whole blood from volunteer donors (22) was processed according to regional American Red Cross Blood procedures. Briefly, whole blood collected into a CP2D blood collection bag was centrifuged (all centrifugations at 20–24 °C) at 2500 × g for 3.5 min. Platelet-rich plasma (PRP) was extracted, and platelets were pelleted at 4300 × g for 6 min. Residual plasma was removed, leaving 45–65 mL for platelet resuspension. The resulting PC was maintained at room temperature (RT) on a rocker, and typically contained 1–2 × 106 platelets μL−1. Following filtration through a Purecell PL leukoreduction filter, platelet purity was 99–99.99%.

PC was prepared in the laboratory from 11 healthy volunteers as described above, with the following variations. Blood was centrifuged in polypropylene tubes at 135 × g for 15 min, and this was followed by pelleting of platelets at 1500 × g for 10 min. PCs were resuspended in one-fifth of the original plasma volume. The time from venipuncture to storage PC was approximately 3 h.

Apheresis platelets from a single blood donor were collected using the AMICUS or Gambro instruments. Briefly, donor whole blood withdrawn continuously was anticoagulated with sodium citrate solution (8:1 v/v) and centrifuged, producing platelet-rich, leukocyte-depleted plasma. This PRP contained approximately 1.5 × 109 platelets per microliter. Contamination was minimal, as described above. The number of units tested was based on availability of product from the blood bank, and obtaining statistically significant and clinically relevant data.

Measurement of CD154 in plasma and platelet lysates

Whole blood-derived platelets from two donors were processed by the PRP method and leukoreduced. Plasma CD154 levels were determined at 4 h, day 1 and day 5 of storage. Platelet pellets were lyzed in Nonidet P-40 lysis buffer containing protease inhibitor cocktail (Sigma, St Louis, MO, USA). CD154 levels were measured with an enzyme-linked immunosorbent assay (ELISA) specific for CD154 developed in our laboratory (Standard – BenderMed Systems: range: 6–400 pg mL−1).

Effects of thrombin and aspirin on platelet activation

Platelet samples were removed from PC at the indicated times, and washed twice in HEPES-modified Tyrodes. Platelets were resuspended in buffer and incubated with agitation at RT for the indicated time with or without the addition of 0.2 U mL−1 thrombin (Sigma-Aldrich, St Louis, MO, USA).

To inhibit COX-1 activity, platelets were washed as described above and pretreated with aspirin (acetylsalicylic acid, 10 μg mL−1; Sigma-Aldrich) for 30 min, and washed twice.

Detection of CD154 on platelets by fluorescence-activated cell sorting (FACS)

Washed platelets were resuspended in HEPES-modified Tyrodes plus 0.5 mg mL−1 bovine serum albumin, incubated with mouse monoclonal IgG1 anti-human CD154 antibody conjugated to fluorescein isothiocyanate (FITC) or FITC-labeled mouse IgG isotype control antibody, washed twice, and analyzed with a Becton Dickinson FACScan flow cytometer. Platelets were identified by forward and side scatter distribution validated by staining with one of two platelet-specific antibodies (anti-CD42a or anti-CD42b). Data analysis was performed using Becton Dickinson's cell quest software.

Preparation of CD154-depleted PC supernatant

PCs stored for 5 days were immunodepleted of CD154 with a monoclonal anti-human CD154 antibody (Calbiochem, La Jolla, CA, USA). PC was centrifuged at 1500 × g for 15 min, and PC supernatant (PCS) was precleared using protein G–Sepharose (Amersham, Piscataway, NJ, USA) for 1 h at 4 °C and incubated for 2 h with monoclonal anti-CD154 cross-linked to protein G–Sepharose beads (Seize-X Immunoprecipitation kit; Amersham). The Sepharose beads–antibody–CD154 complex was removed by centrifugation. The PCS and resulting CD154-depleted PCS (dPCS) was assayed for CD154 levels as previously described.

Fibroblast cell culture and activation

Primary strain human lung fibroblasts were previously isolated in our laboratory and maintained as described [4]. Cells were between passages 4 and 15. Fibroblasts were serum-starved for 24 h prior to addition of PCS or dPCS to a final dilution of 1:50. Neutralizing CD154 and isotype control antibodies (mIgG1) were used at 1 μg mL−1, determined by titration.

Analysis of sCD154, IL-6 and prostaglandin E2 (PGE2)

Stored PC samples were aseptically removed and centrifuged for 15 min at 1500 × g, generating platelet-poor plasma. sCD154 levels were determined by ELISA in triplicate. Fibroblast-free supernatants were harvested (72 h). The IL-6 concentration was determined by ELISA (biotinylated anti-human IL-6 and IL-6 standard; BD Biosciences, San Jose, CA, USA), and the PGE2 concentration was tested using a PGE2 enzyme immunoassay kit (Cayman Chemical Co., Ann Arbor, MI, USA).


Fibroblasts grown in eight-well Permanox chamber slides (Nalge Nunc) at a density of 5 × 104 cells per well with medium (unstimulated), PCS or dPCS (24 h) were washed twice with phosphate-buffered saline plus 0.05% Tween-20, and endogenous peroxidase activity was quenched with 3% H2O2 for 10 min. Cells were blocked with 5% horse serum (1 h at RT), and incubated overnight at 4 °C with 10 μg mL−1 mouse anti-human Cox-2 antibody (Cayman Chemical Co.) or isotype control mouse IgG1 antibody (Caltag Laboratories, Burlingame, CA, USA). Secondary antibody (biotin-labeled anti-mouse IgG, 1:200 dilution; Vector Laboratories, Burlingame, CA, USA) was added at RT for 1 h, followed by streptavidin–horseradish peroxidase (1:1000 dilution for 1 h at RT) (Jackson ImmunoResearch Laboratories, West Grove, PA, USA). Aminoethyl carbazole (Zymed Laboratories, South San Francisco, CA, USA) was used to visualize cell staining.


Where indicated, data were analyzed by one-way analysis of variance and the Student–Newman–Keuls Multiple Comparison Test using graphpad instat software v. 3. Linear regression analysis was performed by Prometheus Research.


Platelets prepared for transfusion are partially activated

Random donor whole blood-derived PCs prepared by the PRP method and stored for 5 days were sampled early (day 1) and late (day 5). FACS analysis of membrane CD154 staining (Fig. 1), gated on CD42a-positive events, revealed that on day 1 of storage, about 27% of the platelets expressed low levels of CD154 (mean fluorescence intensity [MFI] = 22) on their surface; this increased to 60% by day 5 (MFI = 68). This trend was observed in all six samples of PC tested. Moreover, CD154 expression increased following thrombin stimulation of day 5 platelets. The majority of thrombin-stimulated platelets (85%) were positive for CD154 (MFI 86). This indicates that platelets are partially activated during the storage process, and further activated by stimulation with agonist.

Figure 1.

 Platelets prepared by the platelet-rich plasma method express greater amounts of surface CD154 after 5 days of storage. Platelets prepared for transfusion were sampled during their 5 days of clinical storage. The histograms represent CD154 staining (open histograms) gated on CD42a-positive events, and isotype control staining (gray-filled).

To test whether sCD154 was generated in stored PC plasma, 16 random donor PCs were sampled on the day on which they were used for transfusion (day 3, 4 or 5). All PCs tested had substantial amounts of sCD154 (average: 11.3 ng mL−1) as assessed by ELISA (Fig. 2A). We tested three PCs (nos. 17–19) following pre- and postleukocyte reduction filtration. The sCD154 ELISA results (Fig. 2B) indicate that filtration does not remove sCD154 from PCs.

Figure 2.

 Platelet concentrates (PCs) prepared by the platelet-rich plasma method have high levels of sCD154. (A) Platelet-poor plasma from all PCs contained sCD154 [average concentration: 11.3 ng mL−1 (dashed line)]. (B) Three PCs tested prefiltration and postfiltration demonstrated that poststorage leukoreduction of stored PC has no effect on sCD154 level. Each PC supernatant concentration is the mean of triplicate samples; error bars = SD.

To determine the potential total amount of CD154 released in stored platelets, two single donor PCs were prepared in the laboratory and stored under blood bank conditions, as we do not have access to stored PCs prior to day 3. CD154 levels were measured in plasma and platelets by ELISA. sCD154 was present at high levels (721 ng/5 × 1010 platelets) within 24 h in plasma, representing a twenty-onefold increase of sCD154 in the first 20 h of storage (Table 1), and demonstrating that 52% of total CD154 contained in platelets was already expelled. Importantly, levels continued to increase, and by day 5, had reached 838 ng per 5 × 1010 platelets or 62%. Therefore, sCD154 amounts approaching 1 μg are transfused with platelets into patients. Additionally, thrombin treatment of platelets released additional sCD154, 37% and 21% for days 1 and 5, respectively (data not shown). This suggests that a subset of storage platelets remain bioactive for transfusion.

Table 1.   Potential total amount of CD154 in platelets
 ng CD154 / 5×10 e10 platelets
 4hDay 1Day 5
  1. *ND, not done.

  2. CD154 levels were measured (ELISA) for plasma and platelet lysates at the times indicated. The experiment was done for two donors and the results were consistent.

Platelet lysateND*662512
 % in plasma5262

Kinetics of sCD154 release from stored platelets

All random donor PCs contained significant sCD154 levels by the time that they were used for transfusion. Stored PCs prepared in the laboratory were examined twice daily for 5 days to measure the early CD154 release kinetics. CD154 levels rose almost immediately (Fig. 3). A linear regression predicting sCD154 concentration from storage time was highly significant: P < 0.001, R2 = 0.748. Thus, sCD154 concentration increases significantly with time. This finding is important, as platelet transfusions are typically performed on days 3–5 following collection.

Figure 3.

 A maximum concentration of sCD154 in platelet concentrates (PCs) is achieved after 3 days of storage. PCs were sampled twice daily for 5 days to determine sCD154 levels by ELISA. sCD154 levels rose quickly, reaching a maximum at 72 h. Each data point is an average of triplicates; error bars = SD.

Effects of thrombin and aspirin on sCD154 release

Previously, platelet sCD154 release was demonstrated only after agonist stimulation; our data indicate that preparation and/or storage of platelets induces ‘spontaneous’ sCD154 release. To compare sCD154 release with and without thrombin addition, laboratory-prepared PCs were examined by time course and sCD154 release measured by ELISA. Both unstimulated and thrombin-stimulated platelets (Fig. 4) approximately doubled the sCD154 level in the supernatant at 5 min in comparison to unstimulated platelets (t = 0). Unstimulated platelets maintained a constant sCD154 level for approximately 1 h, before the level slowly increased to 1.5 ng mL−1 at 3 h, reaching 2.3 ng mL−1 at 20.5 h. Conversely, thrombin-stimulated platelets had the steepest increase in sCD154 levels between 5 and 15 min of incubation (2.4 ng mL−1), and the level continued to increase, albeit less sharply, reaching 5.7 ng mL−1 at 20.5 h.

Figure 4.

 sCD154 release kinetics are greatly enhanced by addition of thrombin. sCD154 levels increased in both unstimulated and thrombin-stimulated platelet concentrates prepared in the laboratory. Levels in thrombin-stimulated platelets continued to increase significantly, reaching 5.7 ng mL−1 at 20.5 h. Each time point is an average of triplicate samples; error bars = SD.

A multiple linear regression model predicting sCD154 concentration during storage in the presence and absence of thrombin demonstrates that sCD154 concentration increases much faster with thrombin present (P < 0.001). These data support the hypotheses that there are different degrees of platelet activation and that platelets prepared for transfusion are partially activated.

The substantial increase in sCD154 release as a result of thrombin stimulation raised the possibility that thrombin proteolysis activity may be involved in membrane release of sCD154. We tested other platelet agonists for their ability to release sCD154 by ELISA (unpublished data). Thrombin receptor-activating peptide (TRAP) and collagen strongly released sCD154, in an analogous way to thrombin-induced release. This suggests that the proteolytic activity of thrombin has a minimal effect on membrane sCD154 release, as neither agonist has an intrinsic proteolytic function. Others have reported that CD154 is cleaved by a putative metalloproteinase, and a direct inhibitor of this protease abrogated thrombin-induced sCD154 release from platelets [21].

Platelet aggregation involves the synthesis of thromboxane A2 (TxA2), a potent platelet activator and the main platelet product of arachidonic acid metabolism by cyclooxygenase-1 (COX-1). Aspirin, a COX inhibitor, can block this process, and its effects on sCD154 release from stored platelets were examined by pretreatment with or without aspirin, and sampling at various times over 4 days. Fig. 5 demonstrates no significant difference in sCD154 levels or release in aspirin-pretreated platelets as compared to untreated platelets, implying that CD154 release is not related to COX activity and does not require aggregation.

Figure 5.

 Aspirin does not inhibit sCD154 release from platelet concentrates. Platelets were untreated or pretreated with aspirin for 4 days. Pretreatment with aspirin did not significantly alter sCD154 release from stored platelets. Each time point is an average of triplicate samples; error bars = SD.

A linear regression test of the significance of differences in sCD154 concentration during storage with and without aspirin demonstrated no significant effect of aspirin on sCD154 accumulation (P = 0.570). Consistent with our data, Nannizzi-Alaimo et al. [22] found that aspirin was not effective at inhibiting sCD154 release from platelets stimulated with ADP or TRAP, but did reduce the amount of sCD154 released following collagen activation.

sCD154 release from apheresis platelets

Our data thus far concern random donor platelets prepared from whole blood. To determine whether single donor platelets, prepared by apheresis technology, and leukoreduced prestorage, were similarly activated to express CD154 during storage, apheresis products stored under normal blood bank conditions were sampled on days 3–5 to measure membrane and soluble CD154 levels. Owing to infectious disease testing requirements, single donor PCs were not available to us for testing until day 3. Data from one of three single donor PCs tested are presented (Fig. 6), and are representative of the other two apheresis products. The single donor PC had a maximum sCD154 concentration at the first time point (72 h), and this did not change over the two storage days (Fig. 6A). FACS analysis of apheresis platelets stained with anti-CD154 and gated on anti-CD42b-positive events (CD42b is a platelet-specific marker found on > 95% of cells in apheresis PCs) showed that CD154 was already expressed on the membrane of apheresis platelets by day 3 (Fig. 6B). As with sCD154, the levels of membrane CD154 did not change over the next 2 days of storage. We conclude that single donor/apheresis platelets are as equally activated as random donor platelets prepared from whole blood, and translocation of CD154 occurs ‘spontaneously’ in both preparations.

Figure 6.

 Single donor apheresis platelets are partially activated and express CD154. (A) Platelet-free supernatant contained high sCD154 levels by day 3 that did not change with extended storage. Data are the average of triplicate samples; error bars = SD. (B) CD42b-positive platelets were partially activated, expressing CD154 at day 3. No change in CD154 expression was seen at day 5.

CD154 in PCs is biologically active

It is possible that CD154 released by platelets during clinical storage is not biologically active and is therefore not relevant to transfusion complications. We previously showed that fibroblasts activated via CD40 engagement are induced to express COX-2, and subsequently produce large amounts of PGE2, a potent proinflammatory mediator [23], and the key inducer of fever in humans. IL-6 is a proinflammatory cytokine also known to be induced by CD40 ligation [4]. Both mediators could play a role in transfusion reactions.

PC samples were stored for 3–5 days in the blood bank. We incubated CD40-expressing human lung fibroblasts with medium alone (unstimulated), PCS (1:50 dilution) or dPCS (1:50 dilution). After 24 h, fibroblasts were stained with an anti-COX-2 or isotype control antibody (Fig. 7A). Unstimulated fibroblasts expressed low levels of COX-2, whereas COX-2 is potently upregulated in fibroblasts incubated with PCS. CD154 depletion from PCS significantly reduced the intensity of COX-2 staining. Fig. 7B shows that CD154 depletion was measured near the detection limit of our ELISA (0.16 ng mL). It is known that, in vivo, sCD154 is released during thrombotic events, reaching concentrations in the circulation of 5–10 ng mL−1 [24]. Therefore, any remaining CD154 in dPCS is presumably present at concentrations too low to stimulate COX-2 upregulation in fibroblasts, implying that CD154 is an important player in the observed effects.

Figure 7.

 Platelet-derived sCD154 induces COX-2 expression in human lung fibroblasts. (A) Human lung fibroblasts were stimulated with medium (unstimulated), platelet concentrate supernatant (PCS) or CD154-depleted PCS (dPCS). After 24 h, cells were stained with a COX-2-specific antibody or an isotype control. Unstimulated cells express low levels of COX-2. PCS strongly activated lung fibroblasts, upregulating expression of COX-2. The induction of COX-2 was partly dependent on CD154, as dPCS had a significantly reduced capacity for COX-2 induction. (B) sCD154 ELISA results confirmed that high levels in PCS were significantly reduced by immunodepletion (dPCS).

Unstimulated human lung fibroblast supernatants were then compared to stimulated PCS, PCS in the presence of neutralizing anti-CD154 antibody (5c8) or a control antibody to measure PGE2 levels, and PCS and dPCS to measure IL-6 induction. PCS stimulation greatly increased both PGE2 and IL-6 levels in fibroblast supernatants (Figs 8 and 9, respectively). Blocking CD154–CD40 interactions with anti-CD154 antibody inhibited PGE2 production 2-fold, a statistically significant reduction (P < 0.001). The control antibody had no effect (Fig. 8A). IL-6 induction was inhibited to a similar degree when PCS was depleted of CD154 prior to fibroblast stimulation (P < 0.001 for PCS vs. dPCS) (Fig. 9).

Figure 8.

 Platelet-derived sCD154 can activate human lung fibroblasts to make prostaglandin E2 (PGE2). Cell-free supernatants were tested for PGE2 at 72 h. (A) Unstimulated fibroblast supernatants contained 134 pg mL−1 PGE2. Platelet concentrate supernatant (PCS) stimulation induced a significant increase in PGE2 levels (4898 pg mL−1, P < 0.001). Addition of neutralizing CD154 antibody inhibited PGE2 production (2563 pg mL−1) by about 50% compared to PCS stimulation (P < 0.01). The isotype control antibody had no effect. These results represent nine of 11 experiments. (B) Three experiments demonstrate more robust inhibition of PGE2 after blocking of the CD154–CD40 pathway. Unstimulated fibroblasts (821 pg mL−1 PGE2) produced large quantities of PGE2 in the presence of PCS (18535 pg mL−1, P < 0.001). The addition of the neutralizing CD154 antibody completely inhibited this response (548 pg mL−1, P < 0.001). The isotype control had no effect. All data presented represent a minimum of triplicate samples; error bars = 1 SD. *Statistically significant difference in comparison to unstimulated cells. !Statistically significant difference in comparison to PCS stimulation.

Figure 9.

 Platelet-derived sCD154 activates human lung fibroblasts to make interleukin-6 (IL-6). After 72 h, cell-free supernatants were tested for IL-6. (A) Lung fibroblasts potently expressed IL-6 following platelet concentrate supernatant (PCS) stimulation (7720 pg mL−1) in comparison to unstimulated fibroblasts (304 pg mL−1) (P < 0.001). For nine of 11 experiments, depleted PCS (dPCS) stimulation resulted in about 50% inhibition of the IL-6 induction (2786 pg mL−1) in comparison to PCS (P < 0.001). (B) In two other experiments, stimulation of fibroblasts with dPCS completely abrogated induction of IL-6 in comparison to stimulation with PCS (P < 0.001). All data presented represent a minimum of triplicate samples; error bars = 1 SD. *Statistically significant difference in comparison to unstimulated cells. !Statistically significant difference in comparison to PCS stimulation.

Interestingly, we observed variability in the effect of depleting/blocking CD154. Three of 11 experiments demonstrated potent activation of IL-6 and PGE2 by PCS, and almost complete inhibition of PGE2 and IL-6 induction when blocking antibody or dPCS was used (Figs 8B and 9B, respectively). In the majority of cases, the inhibition observed was about 50% in comparison to PCS stimulation (Figs 8A and 9A). We speculate that these differences may result from variability in ‘activation states’ of PCs prepared for transfusion, with release of different types or amounts of proinflammatory mediators.


The present study demonstrates that PCs used for transfusion have high levels of platelet membrane-bound CD154 and sCD154 in stored plasma. This is consistent with the ‘platelet storage lesion’, which refers to platelet activation during storage [25]. We examined PCs generated by the two most common clinical techniques in the USA: the PRP method and apheresis. Apheresis platelets contained high levels of surface CD154 and sCD154 at day 3, and levels remained high through day 5 (Fig. 6). A thorough time-course analysis of CD154 membrane translocation and release in PC prepared by duplication of the PRP method in the laboratory revealed that 27% of CD154 was expressed by platelets on day 1 of storage, with levels increasing to 60% by day 5.

In support of these data, investigation of CD154 release in PC prepared in the laboratory demonstrated that sCD154 is present in stored plasma in amounts greater than or equivalent to 50% of the total CD154 amounts contained in platelets as early as day 1 of storage. Importantly, CD154 release continues, and by day 5, at least 60% of total CD154 is found in plasma, and transfused with platelets into patients. Although we were not able to test blood bank-prepared PCs immediately after processing, CD154 was not found on the platelet surface of freshly prepared PCs, and plasma sCD154 levels were correspondingly low (data not shown).

These data support the hypothesis that both PC and apheresis platelets become activated by collection, storage, or both. Interestingly, Prasad et al. [24] established that sCD154 functions as a platelet agonist, binding the αIIbβ3 integrin to initiate outside-in signaling, resulting in platelet activation. This autocrine loop mechanism of activation represents a pathway by which stored platelets can be induced to express greater amounts of CD154 over time. As most platelets are not transfused until days 3–5, our findings indicate that recipients of single donor/apheresis platelets, and whole blood platelets, receive a substantial and potentially deleterious dose of sCD154.

Additionally, thrombin-stimulated platelets still contain and release sCD154 throughout their 5-day storage period. This is consistent with earlier findings that although stored platelets are partially activated, they can still respond to agonistic stimuli [25] and are still functional in vivo. Indeed, their state of partial activation may be key to their clinical efficacy.

Our data provide evidence that platelet CD154 in stored components is biologically active and potent, as PCS activates human lung fibroblasts to express COX-2, PGE2 and IL-6 (Figs 7–9). This effect was partially blocked by depleting PCS of CD154 or using a neutralizing anti-CD154 antibody, impeding IL-6 and PGE2 induction, respectively. These results correlate with our previous studies demonstrating dramatic COX-2 upregulation and concomitant PGE2 production via CD154–CD40 interaction in human lung fibroblasts using recombinant CD154 [23]. In some cases, PCS-induced PGE2 and IL-6 were completely blocked by inhibition of CD154–CD40 interaction (Fig. 9). Although this result seems discrepant, we recognize that COX-2 expression can be induced by different signaling pathways via other cytokines and lipid mediators. For example, platelet TxA2 release was shown to act in a paracrine fashion to upregulate COX-2 expression in endothelial cells [26]. Thus, these observed differences may be attributed to the variability in potential mediator content among different PC preparations. We speculate that in experiments where PCS activation of fibroblasts appeared to be completely dependent on CD154, the PCs contained levels of other cytokines or lipid mediators that were too low to cause activation on their own.

Our results strongly indicate that stored platelet-derived CD154 is viable and present in concentrations sufficient to induce platelet activation, and potentially capable of eliciting an adverse effect during transfusion. Consistent with our previously reported findings [20], these studies demonstrate high sCD154 levels (average concentration: 11.3 ng mL−1) in 16 different single donor PCs prepared for transfusion, on the day that they were transfused. This is an important finding, because PCs generated by the PRP method are widely used platelet products. Furthermore, we demonstrate that neither prestorage nor poststorage leukoreduction of apheresis platelets or PCs, respectively, had any effect on soluble or surface CD154 levels. Others have also demonstrated that prestorage leukoreduced platelets stored under standard transfusion conditions liberate a variety of mediators (transforming growth factor-β1, PF4, RANTES), including sCD154, over a 5-day period [15,19,27,28]. Furthermore, platelets are the major source of bioactive cytokines following prestorage leukoreduction, independently of whether they were treated photochemically or gamma irradiated [19].

It was previously suggested that FNHTRs result from cytokine accumulation in PCs [15], and our work supports this idea. We propose that sCD154 is an additional ‘cytokine’ player in the pathogenesis of FNHTRs, and transfused sCD154 may lead to inflammatory and vascular leak effects, as seen in TRALI [29,30]. It was suggested that combined leukoreduction and plasma reduction may be required to prevent FNHTRs [31]. Heddle et al. [18] predicted that eliminating cytokines from transfusion products would prevent these reactions; however, clinical trials demonstrated that even with plasma depletion, FNHTRs still occurred in 14.9% of patients, compared to 22.6% of those receiving poststorage leukoreduced platelets. Although plasma reduction eliminates most sCD154 accumulation in platelet bags, stored platelets also express membrane CD154, and transfusion of significant amounts of CD154-positive platelets may account for the remaining transfusion reactions seen in plasma and leukocyte-depleted platelets.

The effect of infusing a large dose of sCD154 may vary with the underlying condition of the recipient. Several pathologic conditions, such as inflammatory bowel disease and Type 1 or 2 diabetes [32,33], have been correlated with increased circulating sCD154 levels. Platelet transfusions in these individuals may exceed the threshold for sCD154 plasma levels. Transfused CD154, in concert with other cytokines and prostaglandins, could result in amplification of biological responses, resulting in the observed adverse clinical events. Screening patients for sCD154 plasma levels could, in theory, identify patients at higher risk for developing transfusion reactions.

In summary, our data indicate that patients receiving platelet transfusions also receive large doses of soluble and membrane-bound CD154. The CD154–CD40 pathway has emerged as a central player in both immunologic and inflammatory processes. Because of the potency of this molecule, we postulate that CD154 may be a contributor to FNHTRs and other immunologic complications following platelet transfusions. Our conclusions are supported by recent studies demonstrating that sCD154 can activate adherent polymorphonuclear leukocytes and is associated with TRALI [34]. Others have provided evidence that CD154 is implicated in adverse platelet transfusion events such as FNHTRs [29,35]. Additional studies in animal models will be required to verify the role of CD154 in transfusion reactions, so that measures can be taken to reduce these unwanted events.

Disclosure of Conflict of Interests

This research was supported by the National Institutes of Health grants DE011390, ES01247, T32AI07285, HL078603 and HL086367.