Aging and death signalling in mature red cells: from basic science to transfusion practice

Introduction

The red blood cells (RBCs) aging process is considered as an issue of special scientific and clinical interest. It represents a total of unidirectional, time-dependent but not-necessarily linear series of molecular events that finally lead to cell clearance¹. Under normal circumstances, all human RBCs live approximately 120±4 days in blood circulation, implying the existence of tightly regulated molecular mechanism(s), responsible for the programming of the lifespan and the nonrandom removal of senescent RBCs^2,3. Although the RBCs have already been used as a model for aging study¹, the molecular participants, as well as the signalling pathways involved, are not yet completely clarified.

RBCs storage under blood bank conditions is far from being considered analogous to the physiologic in vivo aging process. The putative implicated in vivo signalling pathways are expected to be more-or-less preserved under in vitro conditions, nevertheless slightly modulated, in response to a totally different environment. A storage period of up to 35–42 days at 4 °C probably is not a “congenial interlude” of the physiological maturity process and definitely does not represent an ignorable time period, compared to the RBCs lifespan. Stored RBCs age without the normal adjacency of other cells or plasma, which continuously provide them with survival factors and signals and, moreover, they are obligated to share their living space with their own and other cells’ wastes. Since no clearance mechanisms seem to function, senescent RBCs are probably sentenced to “survive” for a longer period than they were probably programmed for. Although there is evidence suggesting that storage disturbs the physiological RBC aging process^4–7, the mechanistic basis of the aging progress inside the blood unit and the functional reactivity of the modified RBCs in vivo, remain still elusive.

Given the fundamental need for safe and efficient transfusions, the clinical impact of stored blood, as a function of the storage parameters, has attracted considerable attention. Clinical trials that focus on the potential adverse clinical consequences of transfusing older storage-age RBCs units vs. younger ones have already been reported⁸. Apart from the skepticism around their design and execution⁹, these studies indicate the necessity of thorough examination of the storage effect on packed RBCs, before analyzing their potential impact on the clinical outcome.

This review focuses on the current knowledge on aging and death signalling pathways operating in both in vivo systems and stored RBCs and suggests future directions in the preservation science, helpful for addressing what seem to be the current critical questions in transfusion medicine.

RBC aging, senescence and in vivo death signalling pathways

RBCs experience a range of continuous metabolic and physical damages as they age, such as membrane vesiculation¹⁰, haemoglobin (Hb) modifications and progressive failure of both, cellular homeostasis and antioxidant defenses. The increase in RBCs density¹¹, the nonenzymatic glycation of Hb¹² and the deamidation of protein 4.1b to 4.1a^13,14 have been widely used as sensitive RBC age markers. In fact, numerous post-translational protein modifications, including phosphorylation, oxidation and aggregation are functionally involved in the regulation of RBCs homeostasis and lifespan.

Despite these cumulative events, the senescent signals, namely the molecular measure of RBCs age, do not seem to gradually express in cells. On the opposite, they appear as a snap, rapid and non-linear cascade of events at the terminal stage of the aging process, probably shortly before RBCs removal by the reticuloendothelial system³. Taken into consideration the inability of mature RBCs to synthesize new proteins, the recognizable markers must derive from modifications in pre-existing molecules. Their generation is most probably the ultimate step of more than one signalling pathways, working in a sophisticated context of molecular interplays. To the current knowledge, the “RBC aging phenotype”, namely the repertoire of age-dependent alterations, can be safely associated with a reported decline in metabolic activity, a progressive cell shape transformation, a membrane remodelling, as well as with oxidative injury, microvesiculation and exposure of surface removal markers. The only common feature of these modifications is that they all, directly or indirectly, trigger erythrophagocytosis¹⁵.

Microvesiculation

Membrane microvesiculation is part of the RBCs maturation. It represents a well regulated process that is accelerated in older cells¹⁰. Depending on the circumstances, it may function against (by contributing to irreversible membrane/Hb loss) or in favour (by carrying away damaged and signalling effective cell components) of the growing RBCs¹⁶. The exocytosis of non-functional proteins and senescence marks through vesiculation not only protects the RBCs from premature death but also indicates that the same recognition signals mediate the rapid removal of old RBCs and vesicles from the circulation¹⁶. Through the continuous release of vesicles, RBCs indices change as they age¹² and cell density progressively increases concomitantly to a decrease in cellular deformability and membrane flexibility. These modifications have been appreciated as major determinants of RBCs premature in vivo removal.

The Band 3-based aging pathway

So far, numerous RBCs aging pathways have been proposed based on cellular changes identified in older RBCs. Among them, the clustering¹⁷ and/or the breakdown of Band 3^18,19 is probably the central step in the major immunologically mediated pathway, leading to the generation of a powerful senescent signal, a senescent-specific neo-antigen, in vivo. Similar processes seem to be responsible for premature RBCs clearance in haemoglobinopathies, membrane protein deficiencies, Down’s syndrome, Alzheimer’s disease etc^17,19. Upwards, free radical oxidation may be an important factor underlying the formation of the senescent signal. There is evidence connecting the RBC oxidation levels in vivo to the breakdown of Band 3 and to the autologous IgG binding²⁰, as well as the elevated membrane-bound modified Hb to the high incidence of autologous IgG binding²¹. Furthermore, formation of advanced glycation end products²², binding of oxidative denatured Hb to Band 3²³ and the following modifications in tyrosine phosphorylation¹⁷, may induce the observed topographic redistribution of Band 3. Downwards, the senescence neoantigen appearance induces the binding of both autologous IgGs and probably C3 fraction of the complement to the membrane, triggering erythrophagocytosis^18,24,25.

Calcium homeostasis

Distorted calcium (Ca²⁺) homeostasis is probably part of another aging-related pathway either as a triggering factor for aging or as its consequence²⁶. Although the path’s regulation under physiological conditions is obscured, membrane-associated Bcl-X_L and Bak, which form functional interactions with survival factors of the plasma, might mediate it². Calcium influx is clearly correlated to oxidative damage, vesiculation, dehydration and deformability defects suffered by the senescent RBCs²⁷. Furthermore, there is an established functional connection between calcium influx and apoptosis-like events in mature erythrocytes^27–30. That calcium-involved set of pathways working in RBCs in response to stress (oxidative, osmotic etc) has been called eryptosis. Two underlying signalling pathways have been reported: (i) formation of prostaglandin E2 that leads to the activation of Ca²⁺-permeable cation channels and (ii) phospholipase A2–mediated release of platelet-activating factor that activates a sphingomyelinase, leading to formation of ceramide. Increased intracellular Ca²⁺ and ceramide levels lead to PS exposure. Moreover, calcium activates Ca²⁺-sensitive K⁺ channels, leading to cellular KCl loss and cell shrinkage. In addition, Ca²⁺ stimulates μ-calpain transglutaminase-2 and, occasionally, caspases that degrade/crosslink the cytoskeleton proteins, resulting in loss of membrane integrity, deformability and blebbing. Finally, Ca²⁺ disrupts the critical interaction between phosphotyrosine phosphatase and Band 3³¹. Eryptosis may be a haemolysis escape mechanism of defective erythrocytes^29,30 but its relevance - if any- to RBC aging process still has to be firmly established by future studies.

Caspase signalling and PS exposure

Erythrocytic procaspase 3 is in vitro activated under oxidative stress, leading to Band 3 modifications, PS exposure and erythrophagocytosis^32,33. Such a mechanism is very likely to have an in vivo physiological role in RBCs aging/clearance, as indicated by the active caspase -3 and -8 detection, as well as by the formation of the Fas-signalling death complex in the lipid raft membrane microdomains of aged RBCs^34,35. Aged RBCs present lower aminophospholipid translocase activity and higher levels of externalized PS, in comparison with younger ones³⁵. Activation of caspase 3 during RBCs senescence, under the stimulus of increased oxidative stress, could cleave³⁴ or modulate³⁶ Band 3, triggering a cascade which leads to RBC removal. Thus, stimulation of caspases in aged or damaged RBCs could induce their phagocytosis in order to prevent haemolysis.

Most RBC-derived vesicles expose PS¹⁶ and a number of haematologic diseases^29,37 or in vitro stressful treatments of human RBCs result in PS exposure²⁹. Interestingly, PS externalization reflects the rate at which biotinylated RBCs remove from circulation in vivo³⁷. However, in vivo PS exposure in healthy individuals senescent RBCs is still a matter of debate, mainly because of the scepticism against the techniques used to isolate old RBCs^16,38. Notably, Bratosin et al. have recently detected active caspases in a fraction of PS-exposing senescent RBCs, isolated from blood circulation³⁹. Considering the powerful thrombogenic effect of externalized PS, it is more likely that it plays a role in the senescence and removal of normal or stressed RBCs, under certain circumstances, through the activation of different signalling pathways. Nevertheless, this postulation should be revaluated by more sophisticated technical skills.

Other mechanisms of RBC in vivo aging

Other mechanisms that have been proposed in order to explain the selective recognition of old RBCs by macrophages, include the time-dependent desialylation of membrane proteins, which leads to the exposure of “senescence factor glycopeptides” on senescent cells¹. Apart from the mechanism of RBC-bound opsonins, sialic acids and membrane constituents such as CD47, at least in animal models, are suggested to play a further regulative role in the elimination of senescent RBCs, by inhibiting erythrophagocytosis^40,41. Their normal RBC membrane topology/stoichiometry functions as effective “non-eat-me” signals for macrophages bearing their respective receptors. However, such a role for CD47 is probably species-specific²⁸, since there is so far no evidence that Rhesus-null human RBCs with reduced CD47 exhibit increased rate of phagocytosis⁴².

Role of the oxidative stress

The RBCs lifespan dependence on an adequate oxidative stress response, imposed by human diseases has been previously established⁴³. The construction of protein-protein interaction networks in the RBC interactome confirmed that RBCs likely suffer of exacerbated oxidative stress and continuously strive against protein and cytoskeletal damage, recruiting a number of alternative pathways related to protein repair, vesiculation or apoptosis^44,45. Although it currently constitutes an active topic of research, accumulative data suggest a central position for oxidative stress in the RBCs aging signalling. Apart from its impact on Band 3-derived neo-antigen formation and the activation of pro-apoptotic components, oxidative stress also affects Hb and its interactions with membrane components^21,27 and caspase-3⁴⁶. Although the major feature of the oxidatively distorted RBC is the binding of oxidized Hb to high affinity sites on Band 3, the irreducible complexation of Hb with spectrin is also a prominent and probably prior marker of in vivo RBCs aging process, tightly correlated with increased RBC rigidity, decreased deformability, echinocytosis and erythrophagocytosis^47–50. This complex might well promote structural modifications in Band 3 by disturbing the cohesion of the cytoskeleton to the bilayer. Moreover, its formation may threaten the normal assembly of the spectrin tetramer and the phospholipid oxidation via a Ca²⁺-promoted quasi-lipoxigenase activity of the oxidized Hb, leading to PS exposure and signalling recognition by the CD36 macrophage receptor²⁷.

In vivo aging remarks

At the present, it is difficult to assign the relative contribution of each of the various speculative mechanisms to aging and removal of senescent RBCs, but it is likely that all of them play a role in this, apparently complex and tightly regulated, process¹. The new era of RBC proteomics revealed an incredible array of RBC proteins involved in intracellular signalling cascades^51,52, while the available refined models of RBC membrane organization suggest the involvement of other molecules or mechanisms in the aging process. The newly discovered metabolon of Band 3 and glycolytic enzymes complex^53,54, the macrocomplex of 4.1R, which may contribute to the remodelling of RBC surface⁵⁵, the Band 3-to-skeleton bridge of adducin with its possible role in the membrane mechanics and vesiculation⁵⁶ and the proposed structural role of glucose transporter-1 in RBCs⁵⁷ (the membrane expression of which is augmented under RBCs storage in blood banks⁴), are some examples of well organized -by multiple signalling pathways- elements. Once again, the “whether, how and when” of these factors into the aging process deserve to be the object of future endeavours.

Aging, senescence and death signalling pathways in stored RBCs

RBC storage lesion

Biochemical and biomechanical changes in RBCs function and integrity during storage, which affect in vivo survival and function, can be summarized by the term “RBC storage lesion”. Easily recognizable biochemical storage effects are the reduction of adenosine triphosphate (ATP), 2,3-diphosphoglycerate (2,3-DPG), pH and glycolysis rate, the accumulation of lactic acid and the increase in Ca²⁺ intracellularly. 2,3-DPG depletion leads to increased oxygen affinity but after transfusion the 2,3-DPG levels are restored in RBCs. ATP depletion follows the reduction in glycolysis rate, leads to further energetic compromise and is associated -directly or secondary- with a series of biophysical alterations, including cellular shape, membrane stability/deformability and vesiculation⁷. Although membrane deformability has been correlated with RBCs viability after transfusion⁵⁸, there are no direct mechanical fragility measurements as a storage period function. Moreover, the contribution of ATP depletion into the storage lesion and the post-transfusion survival of RBCs have been challenged. The main biophysical effect of storage is probably the loss of membrane and Hb through the progressively increased vesiculation (Figure 1) and the subsequent changes in RBCs mechanical and rheological properties⁵⁹. On the basis of thrombogenetic⁷ and nitric oxide scavenging⁶⁰ potential of vesicles, their transfusion is thought to be connected with adverse clinical outcomes.

Figure 1.

Conventional electron microscopy of RBCs and vesicles released after prolonged storage (35 days) in CPD-A. (A) Membrane blebbing in a RBC that has undergone echinocytic (arrows) transformation. (B) Ultrastructural appearance of a vesicle preparation. Insert: the sediment of the vesicles (v) collected at the final step of the isolation protocol (bars, 0.5 μm).

The release of leukocyte-associated enzymes, cytokines and oxygen radicals have been associated with the storage lesion. Comparative proteomic analysis in stored samples showed predominant accumulation of several bioactive proteins in the supernatant of non-leukofiltered units⁶¹. Furthermore, the oxygen-dependent metabolic modulation is progressively altered during storage and is strongly associated with modifications in Band 3 protein⁶². In fact, the storage results in remarkable RBC membrane remodelling and vesicular protein variability^4,63,64. Comparative analysis of RBC membranes during the progress of storage revealed changes in the presence/amount of proteasomes, chaperones, proteases, kinases, and phosphatases at the membrane^4,63. Growing evidence portrays a time-dependent oxidative assault to Hb, membrane and cytoskeleton components, indicating that oxidative injury is a key part of the physiology of stored RBCs^5,64–67. Interestingly, the “onset” of the storage impact in the membrane-cytoskeleton network is detected in an earlier time-point than previously appreciated^4,5,63,66,68, pointing out that any optimization attempt should be applied in the first weeks of the storage period. Moreover, recent studies showed that aged, stored RBCs have reduced ability to produce nitric oxide, while they suggested that reduction in nitric oxide bioavailability at the endothelium -via the reactivity of the cell-free Hb in stored blood- may underlie the storage lesion. The later may be connected to post-transfusion pathological outcomes, such as microvascular vasoconstriction, platelet activation and pro-oxidant and pro-inflammatory effects⁶⁰.

Band 3-related aging machinery and membrane vesiculation during storage

As expected, some profile differences between in vivo and ex vivo aging have been noticed, including the reverse change in Mean Corpuscular Volume (MCV) index⁷ and the variation in size and shape of the released vesicles. Nevertheless, stored RBCs progressively express some of the typical marks of senescence and erythrophagocytosis^6,15. More importantly, all the hallmarks of the Band 3-related aging machinery have been documented in stored RBCs: early and progressively increased accumulation of oxidized/denatured Hb to the membrane and the cytoskeleton, early complication of spectrin with Hb, aggregation of Band 3 at the membrane level and IgG deposition^4,63,66, while removal of transfused RBCs was remarkably delayed after complement aphaeresis during storage⁶⁹. The storage-related vesiculation is a raft-based process that is exacerbated over time in the cold^7,59. Analysis of the vesicles documented the presence of processed/aggregated Band 3 and denatured/oxidized Hb, IgGs and complement components^4,5,64, verifying the beneficial role of vesiculation in the survival of stored RBCs. It should be noted that the lasting coexistence of vesicles with their cells of origin represents another “novelty” of the ex vivo storage system, the consequences of which to the RBCs are still obscure.

Eryptosis/apoptosis signalling in stored RBCs

The vesicles shed by the stored RBCs expose PS^6,7 but there is still no clear consensus whether this also applies for the RBCs per se^28,70–72. This matter is probably analogous to the previously stated argumentation of in vivo aging, further augmented in ex vivo conditions, because of the limited storage duration and the smaller RBC population that contains proportionally less prone to clearance cells. Notably, the storage-dependent remodelling of the RBC membrane includes the Ca²⁺-promoted binding of sorcin and synexin⁶³. Moreover, prolonged storage has been associated with modifications in Fas-associated proteins and caspase activation in both, RBCs^5,63 and vesicles⁵. This element is significant, considering that WBC-derived functional soluble Fas ligand has been detected in the RBC supernatants after prolonged storage⁷³. Furthermore, semi-quantitative proteomic analysis verified the differences in the expression of lipid rafts-associated proteins between RBC membrane and vesicles^4,7, suggesting that alterations in membrane lipid organization are involved in vesicle formation and in storage-associated increase in PS exposure⁶.

The effect of RBC membrane remodelling³⁹ and Hb oxidation⁴⁶ on caspase activation has already been suggested. Until today, it is not clear whether caspase activation is related or not to apoptosis/eryptosis in stored RBCs. Interestingly, their stimulation coincides with the detection of membrane/cytoskeleton modifications in stored RBCs⁵. Those presumptive caspase-induced membrane modifications might interfere with the mechanical properties of the membrane⁷⁴, the extent of vesiculation⁷ and the signalling of the aging per se⁷⁵. The apoptotic death of RBCs during storage might work as a beneficial alternative to haemolysis, provided that apoptosis is not exacerbated by the storage. Otherwise, apoptosis of RBCs inside the storage bag will be a direct measure for in vitro age/stress-related RBCs changes, with a clear consequence on RBCs recovery and survival, following RBCs transfusion.

CD47-related phagocytosis of stored RBCs

Regarding other signalling pathways, loss of CD47 markers has been documented in RBCs, especially in the older ones⁷¹, during the storage period^4,64,72,76. Apparently, this modification may render RBCs more susceptible to clearance when transfused, but the above association has to be firmly established. CD47 has also been detected in the released vesicles^4,64, showing an ongoing interplay among their various phagocytosis-related signals.

Oxidative stress relevance to the RBC ex vivo aging

As storage progresses, antioxidant defence and oxidative injures on RBC membrane, cytoskeleton and cytoplasm components are getting worse^5,65,66,77. Oxidative mechanisms that lead to normal in vivo aging (see above) are currently considered as the underlying contributor to storage lesions⁶⁷ and are probably related to accelerated and/or aberrant aging of stored RBCs. There is evidence that oxidative stress plays significant role in storage-related vesiculation⁷⁸ and in protein degradation, especially of cytoskeleton proteins⁶⁸. The oxidative state of Hb and the incidence of Hb-induced membrane damage are modulated as a function of the storage period, signifying the key role of Hb oxidation not only in the physiology of stored RBCs but also in the progression of irreversible signalling mechanisms^5,64,66,67. All these factors might contribute to trigger the formation of neoantigens in blood units. Although the direct link between the oxidative stress response and the RBC senescence/removal in stored RBCs remains to be firmly established, there is evidence that RBCs stored under the antioxidant and membrane-stabilizing effect of mannitol exhibit a different expression pattern of senescence marks⁵. In the same context, the storage-related variation in proteasome or protein repair molecules^4,5, likely represents responsive mechanisms against the aging-related assault in major RBC membrane proteins. In the absence of clearance mechanisms inside the blood bag, the final effect of the above functional mechanisms might be multiplied in the recipients, especially in massively transfused ones.

Concluding remarks

Under the continuous pressure of raised clinical questions concerning the safety and efficacy of stored RBCs, there is a strong need for a more thorough and updated investigation of the RBCs storage issue. Clinical practice demands for recruitment of novel biomarkers, as accurate in vitro predictors, not only for RBC in vivo survival but also for functional capability and effects. The proteomic and in silico approaches are valuable for performing a global screening of the storage- and aging- related RBC modifications, their course during the storage period and the impact of storage variations on it. The currently provided data on RBCs aging indicate that RBCs are highly dynamic blood components and provide the basis for further proteomic experiments with a direct impact on transfusion medicine. The future research efforts in RBCs preservation science should be enriched with emerging principles, techniques and knowledge regarding the RBCs interactome, the formation of lipid rafts and other functionally important multiprotein complexes, the repair/destroy mechanisms, the response to oxidative stress, the post-translational processing, the protein sorting into the vesicles and the membrane as a place of execution of the senescence-related apoptosis-like events. This advanced, sophisticated approach would benefit the understanding of the mechanisms that define life and death of RBCs in vivo, in the plastic bags and in the circulation after transfusion. It also represents the safer way to the successful optimization of RBCs preservation protocols and transfused RBCs.

We very much regret that many important studies which made significant contributions to red cell research could not be cited or discussed in this review article due to strict space limitations.