Serotonin Is a Key Factor for Mouse Red Blood Cell Survival

Serotonin (5-HT) is a monoamine originally purified from blood as a vasoactive agent. In nonneuronal tissues, its presence is linked with the expression of tryptophan hydroxylase 1 (TPH1) that catalyzes the rate-limiting step of its synthesis. Targeted disruption in mice of the TPH1 gene results in very low levels of circulating 5-HT. Previous analysis of the TPH1 knockout (TPH1−/−) mouse revealed that they develop a phenotype of macrocytic anemia with a reduced half-life of their circulating red blood cells (RBC). In this study, to establish whether the observed reduced half-life of TPH1−/− RBC is an intrinsic or an extrinsic characteristic, we compared their survival to RBC isolated from wild-type mice. Both in vivo and in vitro data converge to demonstrate an extrinsic protective effect of 5-HT since presence of 5-HT in the RBC environment protects RBC from senescence. The protective effect played by 5-HT is not mediated through activation of a classical pharmacological pathway as no 5-HT receptors were detected on isolated RBC. Rather, 5-HT acts as an effective antioxidant since reduction of 5-HT circulating levels are associated with a decrease in the plasma antioxidant capacity. We further demonstrate a link between oxidation and the removal of damaged RBC following transfusion, as supplementation with 5-HT improves RBC post-transfusion survival in a mouse model of blood banking.


Introduction
Serotonin (5-hydroxytryptamine or 5-HT) is a monoamine originally purified from blood as a vasoactive agent and known for its role in the coagulation process. Since the identification of 5-HT, its role as a neurotransmitter in the central nervous system has attracted considerable attention; yet, increasing evidence substantiates the role of 5-HT in the regulation of important nonneuronal functions. Several receptors, encoded by at least 15 different genes, which are divided into seven subfamilies (5-HT 1-7 ), are activated by 5-HT and are responsible for its effects [1]. Availability of 5-HT depends on the expression of the enzyme tryptophan hydroxylase (TPH), which catalyzes the first and ratelimiting step in its biosynthesis [2]. TPH is found only in 5-HTproducing cells, and exists in two forms, one of which is found predominantly in the central nervous system (TPH2) while the other (TPH1) is in peripheral tissues [3,4]. We and others previously showed that targeted disruption in mice of the TPH1 gene results in very low levels of circulating 5-HT, but normal levels of 5-HT in the brain [3,4]. In depth analysis of the TPH1 knockout mouse revealed that they develop a phenotype of macrocytic anemia resulting from both an ineffective erythropoiesis in the bone marrow and a reduced half-life of their circulating red blood cells (RBC) [5].
In humans, the standard maximum duration of routine RBC storage as approved by the US and European legislation is of 42 days. RBC storage solutions are minimal in their composition, and were mostly designed to maintain RBC in an adequate metabolic state. Indeed, it has long been known that aging of RBC in storage solutions is associated with an increase of potassium, lactate, free hemoglobin levels, and a decrease of glucose, 2,3-disphosphoglycerate (2,3-DPG) and ATP levels [6]. These solutions, however, do not completely protect the banked RBC as they undergo significant morphological and molecular changes during storage. These changes, collectively termed ''RBC storage lesions'', can be catalyzed by oxidation, and include a gradual degradation of membrane proteins such as Band-3 [7][8][9][10][11][12]. Accelerated senescence due to oxidation is not limited to storage conditions as it may also be at play during the normal aging process in vivo. Hence, compromised protection from oxidative damage or accumulation of intracellular reactive oxygen species (ROS) results in a shortened lifespan of RBC and may lead to anemia [13][14][15][16][17]. Consistent with this, RBC contain a strong arsenal of antioxidant enzymes that protect the cells against ROS, including catalase, superoxide dismutase and glutathione peroxidase [18], and a number of circulating antioxidant small molecules such as vitamin E, b-carotene and glutathione, which are present in plasma and may contribute to the antioxidant defense of the organism [19,20].
In this report, knowing that 5-HT is a physiological molecule present in circulation, we show that 5-HT acts as an antioxidant and is critical for RBC survival in mouse. In addition, we provide evidence for a link between oxidation and the removal of damaged RBC following transfusion as supplementation of RBC storage solution with 5-HT improves post-transfusion RBC survival in a mouse model of blood banking.

Mouse transfer experiments
Targeted mutagenesis of the TPH1 gene was previously described [3]. Animals on a C57BL/6J background and 6-8 weeks of age were used. Donor blood was biotinylated by intravenous injection of 50 mg/kg of (+)-biotin-N-hydroxysuccinimide ester (Sigma), and collected with EDTA-coated pipettes. Biotinylated RBC were purified using Histopaque 1083 (Sigma) and transfused into WT or TPH1 2/2 recipient mice on day 0. Percentage of biotinylated cells in blood of recipient mice was determined by flow cytometry (FACS).

Mouse in vitro experiments
Blood was obtained by orbital enucleation and collected in an EDTA tube. RBC were purified using Histopaque 1083. The hematocrit of each sample was adjusted to 5% by adding RPMI supplemented with 20 mM hepes (final osmolality of 315mOsm/ Kg), using an electronic hematology particle counter (Melet Schloesing Laboratories), and stored in sealed 500 ml tubes at 4uC until analysis. Hemolysis was evaluated using Drabkin's reagent, and phosphatidylserine exposure was measured by FACS using APC-annexin V. Tryptophan, 5-hydroxytryptophan, 5-HT, melatonin, trolox and BIMU-8 were from Sigma. The 5-HT receptor agonists, 8-OH DPAT, GR 46611, PNU 22394, WAY 208466 and LP-12 were from Tocris Bioscience. Binding experiments were performed as previously described [21].

Total antioxidant capacity of mouse plasma
Mouse blood was obtained by cardiac puncture in a heparinized tube and centrifuged for 5 min at 500 g. Plasmas were submitted to the ''total antioxidant status'' test following the manufacturer's instructions (Randox Laboratories).

Mouse model of blood banking and transfusion
Blood was collected in a CPDA-1 solution (14%) and leukoreduced using a neonatal high-efficiency leukoreduction filter for RBC (Pall). Blood was transferred into sealed 500 ml tubes, supplemented with 0, 10, 30 or 100 mM of 5-HT and stored at 4uC. At regular intervals, stored RBC were stained with carboxyfluorescein diacetate succinimidyl ester (CFDA-SE), washed twice with PBS and transfused into recipient mice. Recipient mice were bled 5 min (max) and 24h (residual) posttransfusion, and the percentage of donor CFDA-SE + RBC was measured by FACS. The 24h post-transfusion survival was determined by the ratio of the percentage of CFDA-SE + RBC at 24h divided by the percentage of CFDA-SE + RBC at 5 minutes.

Ethics Statement
All animal experiments were conducted in accordance with the French regulations and the European Union guidelines for the care and use of animals. Protocols were approved by the Direction de Services Vétérinaires, the competent French veterinarian agency under agreement B75-15-15.

Results
Previous in vivo data demonstrated that RBC from 5-HTdeficient mice were more sensitive to macrophage phagocytosis and had a shortened in vivo half-life [5]. To establish whether the observed reduced half-life in 5-HT-deficient mice (TPH1 2/2 ) is an intrinsic characteristic, we compared in vitro survival of RBC isolated from TPH1 2/2 as compared to wild-type (WT) mice. RBC were incubated at 4uC and their survival was assessed during a 14 day-period by measuring hemolysis. As shown in Figure 1A, the RBC survival curve was comparable for both genotypes with a half-life of 7.2 days. We thus hypothesized that 5-HT exerts an extrinsic protective effect on RBC and to demonstrate the effect, we first used an in vitro system in which WT mouse RBC were incubated in a medium supplemented with increasing doses of 5-HT. A striking dose-dependent protective effect of 5-HT on mouse RBC was observed ( Figure 1B) as 5-HT supplementation extended the half-life of RBC from 10.7 days (10 mM) to 16.4 days (30 mM), and 29.1 days at the highest dose used (100 mM). Next, we assessed the appearance of phosphatidylserine (PS) at the surface of RBC, as measured by annexin V binding to non-hemolysed RBC of normal forward-scatter and side-scatter characteristics ( Figure 1C). Figure 1C illustrates that increasing doses of 5-HT delayed the appearance of PS at the surface of RBC, demonstrating that 5-HT supplementation prevented the appearance of this senescence marker. In light of these in vitro data, we hypothesized that the decreased in vivo half-life of RBC from 5-HT-deficient mice was due to a suboptimal environment resulting from the lack of 5-HT. We thus performed in vivo experiments in which RBC were transferred in an environment with very low levels of 5-HT ( Figure  2 left panel). More precisely, WT mouse RBC were isolated, labeled with biotin and their survival rate were compared following transfusion either to a WT or a 5-HT-deficient recipient. Figure 2 (right panel) indicates that WT RBC circulating in a 5-HT-deficient environment in vivo have a diminished half-life when compared to RBC transferred to a WT recipient with normal physiological 5-HT level (18.8 days vs 23.0 days). This in vivo result is in agreement with our in vitro data, and demonstrates that 5-HT protects RBC through an extrinsic effect.
Classically, 5-HT exerts its effect by binding to specific membrane receptors and to determine whether the 5-HT protective effect on RBC is receptor-mediated, we conducted a thorough analysis of the 5-HT receptor repertoire expressed by mouse RBC using radioactive binding assays. To insure the purity of the studied material, RBC were sorted electronically, and Ter119 + /CD41 2 cells were selected to specifically eliminate platelets contaminants, known to express some 5-HT receptor subtypes (5-HT 2A [22] and 5-HT 1 -like [23]). None of the molecules tested could bind RBC membranes, showing that mouse RBC do not express 5-HT receptors (Table 1). Moreover, in vitro incubation of mouse RBC in the presence of different 5-HT receptor agonists and 5-HT precursors did not result in a protective effect (Figure 3). These data indicate that the protective effect is not mediated through activation of the classical serotonergic pharmacological pathway.
An alternative mechanism for the 5-HT protective action might be through its antioxidant properties. Indeed, 5-HT has been suggested to be a potent antioxidant in different in vitro tests [24]. To test this possibility, we supplemented mouse RBC incubated in vitro with trolox, a vitamin E analog and well-known antioxidant, or with melatonin, a monoamine derived from 5-HT with recognized antioxidant properties. Our results show that trolox was as effective as 5-HT to prevent hemolysis ( Figure 4A). Addition of melatonin also delayed hemolysis, although a high dose was necessary to observe any protective effect. These data confirm that oxidative damage is an important cause of hemolysis in our model. To further attest that circulating 5-HT acts as an antioxidant in vivo, we compared the total antioxidant capacity of WT vs TPH1 2/2 plasma, the latter containing very low levels of 5-HT. As expected, the antioxidant capacity of TPH1 2/2 plasma was significantly reduced (p,0.0001; Figure 4B). Altogether, the data support the view that 5-HT contributes significantly to the antioxidant capacity of mouse plasma and protects RBC from oxidant-induced senescence both in vitro and in vivo.
Oxidation has been identified as a major cause of the hypothermic storage lesions during blood banking [7][8][9]11]. The likelihood that 5-HT exerts the same protective effect on human RBC in conventional blood bank storage conditions would be of major significance, both for clinical applications in transfusion medicine and for the elaboration of new storage solutions. When blood is donated, it is stored according to defined regulations. One important criterion used by the Regulatory Agencies before approval of a RBC storage system, is a measure of post-transfusion survival that is a measure of in vivo viability, which reflects the quality of stored RBC [25]. Recently, inventive efforts to study the post-transfusion survival of RBC in model animal systems have revealed important information related to the storage lesion and transfusion-related reactions [26][27][28]. Using such an established mouse model of blood banking [27,28] we evaluated the impact of 5-HT supplementation on the 24h post-transfusion survival of  stored RBC. In order to closely mimic the aging process undergone by human RBC during hypothermic storage, mouse RBC were collected in the conventional human storage solution CPDA-1, stored at 4uC after removal of the leukocytes by filtration, stained with CFDA-SE and transfused to a recipient to follow their post-transfusion survival. Results shown in Figure 5 are presented as the percentage of transfused RBC still in circulation 24h after transfusion, following storage at 4uC. Supplementation of CPDA-1 with 5-HT led to a dose-dependent increase in the 24h post-transfusion survival of mouse RBC. Following 7 days of storage, the 24h post-transfusion survival was increased by 26% when stored in the presence of 10 mM 5-HT and by more than 50% with 30 mM and 100 mM 5-HT. These data further show the protective effect of 5-HT and support the view that 5-HT extends the viability of stored RBC by preventing oxidation-related storage lesions.

Discussion
The mouse data presented here reveal that 5-HT acts as a circulating antioxidant to protect RBC from senescence. This protection is effective in vitro and in vivo during aging of RBC as addition of 5-HT to stored RBC extends their shelf life, while reduction of 5-HT circulating levels decreases both the antioxidant capacity of plasma and the RBC lifespan. In addition, using a mouse model of blood banking, we show that 5-HT supplementation to stored RBC improves their post-transfusion survival.
The reduced half-life of circulating RBC previously observed in TPH1 2/2 mice [5] is not due to an intrinsic defect of RBC, but to the reduced level of circulating 5-HT. Indeed, even though TPH1 2/2 RBC expose more PS and are more sensitive to macrophage phagocytosis [5], they have a hemolysis curve comparable to the one of WT RBC when isolated in vitro. More importantly, the transfer of WT RBC to a TPH1 2/2 recipient reduces their in vivo half-life, showing that circulation in a 5-HTdeficient environment reduces their lifespan. Also, 5-HT addition to a RBC storage medium in vitro delays hemolysis and the     [29], it is superior by one log to its detected level of 15.3 nM in platelet poor plasma (6nM in TPH1 KO). 5-HT of the platelet poor plasma has always been considered the ''biologically active'' compartment since most of blood 5-HT is stored in dense granules of the platelet compartment. At this moment, we cannot exclude that part of the reduced survival of RBC in the 5-HT-deficient environment could be due to a macrophage phenotype of the TPH1 KO mouse. Along that line, it was recently shown that 5-HT could modulate macrophage polarization in vitro [30]. Further experiments would be needed to elucidate this possibility but, nevertheless, the decreased RBC half-life and decreased circulating 5-HT level in vivo is in agreement with our in vitro results showing a protective role of 5-HT on mouse RBC. The protective effect played by 5-HT is not mediated through activation of a classical pharmacological pathway as no 5-HT receptors were detected on isolated RBC and no 5-HT agonists could mimic its effect. The protective effect of 5-HT could, however, be replicated by trolox, a vitamin E analog, suggesting that the antioxidant capacity of 5-HT is involved. In fact, 5-HT was previously suggested to prevent lipid peroxidation of platelet membranes in vitro [31]. Similarly, the hydroxyl group of trolox or vitamin E acts as a potent peroxyl radical scavenger to prevent the propagation of free radicals in membranes and plasma lipoproteins [32,33].
Numerous natural substances are recognized for their efficiency as oxidant scavengers in fruits and vegetables. These natural compounds usually have a chemical structure encompassing reductive properties: thiol (glutathione) or phenol functions (vitamin E, resveratrol), flavonoids (quercetin), but also heterocyclic structures bearing the indole nucleus such as tryptophan derivatives [34]. In this capacity, 5-HT is synthesized in rice leaves upon senescence, which is delayed through the antioxidant activity of 5-HT [35]. Indeed, comparison of transgenic rice plants producing either high or low levels of 5-HT showed a phenotype of delayed senescence in the 5-HT-rich plant but accelerated senescence in the 5-HT-poor variety [35]. However, such an antioxidant role for 5-HT has never been described in the animal kingdom as most of the research toward indole-based molecule has been oriented toward the study of melatonin [36]. Our demon-stration that the total antioxidant capacity of TPH1 2/2 plasma was reduced when compared to WT shows that 5-HT significantly contributes to the antioxidant potential of plasma. Oxidative stress is an important factor contributing to reduction of the RBC lifespan, and presence of intracellular and extracellular antioxidants is necessary to prevent oxidative damage. For example, deletion of superoxide dismutase 2 or inactivation of antioxidant enzymes containing a peroxidatic cysteine (peroxiredoxin I and II) causes hemolytic anemia [14,17,37]. Likewise, the half-life of circulating RBC is reduced in the Foxo3 knockout mice, due to a diminished expression of ROS scavenging molecules [16]. More importantly, clinical deficiency of vitamin E in humans can cause varying degrees of hemolysis. For instance, pediatric patients with cystic fibrosis and vitamin E deficiency develop hemolytic anemia [19,20].
On the basis of the data presented, there may be posttransfusion RBC survival advantages in humans when supplementing RBC concentrates with an antioxidant such as 5-HT. This is supported by the fact that 5-HT supplementation improves post-transfusion survival in our mouse model of blood banking. In this established model [27,28], mouse RBC are directly collected in the conventional human storage medium CPDA-1 and leukoreduced using conventional filters, allowing for storage conditions similar to those used during conventional blood banking. This result is crucial since it confirms that 5-HT is not only protecting RBC from hemolysis in vitro, but that it allows survival in the more challenging environment of the in vivo circulation. The measure of the post-transfusion survival is thus an assessment of all storage lesions related to the RBC rheology that occur during hypothermic storage. Also, an important requirement for the approval of new storage solutions is based on an in vivo survival rate of the transfused cells of more than 75%, at 24 hours after transfusion, at the end of the RBC storage shelf life [38,39].
The oxidative injury occurring during storage is thought to be responsible for the progressive decline in deformability of the stored RBC and gradual degradation or aggregation of membrane proteins such as Band-3 [7][8][9][10]40,41]. Both these modifications can lead to the removal of the transfused cells shortly after transfusion, thus reducing the efficiency of the process and putting the patient at risk. Indeed, there is an ongoing debate in the transfusion community as the results of some clinical studies confirm the link between RBC storage lesions (transfusion of old RBC) and the negative clinical outcome of transfused patients [42][43][44][45][46][47]. Thus, an approach to counteract the oxidative stress would improve the quality of stored RBC, extend storage time, and potentially reduce transfusion-related adverse reactions. Along this line, Yoshida's group was able to diminish storage lesions by eliminating oxygen from the storage bag [48,49]. However, it was recently shown that anaerobic storage could impair the RBC metabolic capacity to cope with oxidative stress [50]. Hence, supplementation with an antioxidant (such as 5-HT) would be an attractive strategy to slow the oxidative-accelerated aging process observed during hypothermic storage. Still, care should be taken in the choice of the supplemented antioxidant to avoid any side effect associated with its transfusion to a patient. Taking in consideration the increasing number of prescribed pro-serotonergic agents, transfusion of 5-HT may trigger a serotonin syndrome [51] and, as such, identification of an antioxidant with pre-transfusion protective properties and without post-transfusion risks should be the goal of future research. Knowing the advantages and limitations of the mouse model [52], the beneficial effects of antioxidant supplementation to storage solutions on post-transfusion RBC survival demonstrated in mouse will need to be replicated in humans.