How has proteomics informed transfusion biology so far?

1.1. Transfusion medicine and proteomics: an introduction 

Recent strides in analytical strategies have made hematology evolve from a descriptive medical discipline based on microscopic evaluation of red blood cells (RBCs), leukocytes, and platelets (PLTs), towards a dynamic science at the crossroads of genomics and proteomics [14]. PCR-based analyses have radically changed the study of chromosomal translocation products [15]. New technologies, such as DNA microarray allow a direct analysis of the trascriptome, the mRNA pool, which is the intermediate product of gene expression [16]. However, changes in the expression pattern at the mRNA level do not necessarily correlate with changes at the protein level [17], [18], [19], [20]. It is worthwhile to recall that among the most relevant blood components for transfusion purposes there are RBCs and PLTs, which are enucleated, thus they lack of a proper genome, although they inherit meager amounts of mRNAs from their nucleated precursors. Therefore, proteomic investigations in hematology and transfusion medicine (TM) have been lately attracting a great deal of attention [14], [21], [22], [23], [24], [25], [26], [27], [28].

Proteomics tries to determine the whole protein profile of a specific sample under analysis. Proteomics analyses of blood and blood components definitely represent a challenging task. Blood and blood components display an extremely rich spectrum of proteins, which are involved in the most different activities (coagulation, transport, immune system, cell signaling), as well as by-products of cellular damage and proteins from other tissues [25]. Most importantly, the range of protein concentrations in blood plasma spans from picogram to milligram quantities per milliliter (a dynamic range of 10 orders of magnitude) [29]. At present, no technology exists to simultaneously study proteins throughout this entire dynamic spread. For example, erythrocytes contain large amounts of hemoglobin while more than 90% of the plasma proteome is represented by less than 10 different proteins, albumin being the most abundant [29]. There is an elevated risk of low-abundant protein loss, which ultimately hinders detection of a whole “hidden proteome”. Several approaches have been proposed with the goal to either reduce sample complexity – by splitting protein fractions (electrophoresis pre-fractionation [30]) – or to lower the “analytical noise”, through the removal of high-abundant species. As far as the latter approach is concerned, immunoaffinity depletion [31], [32], alone or in combination with electrophoretic pre-fractionation [33], and combinatorial ligand libraries [34] (Fig. 1) have been recently gaining momentum. Nevertheless, the removal of high-abundant proteins has some considerable detrimental pitfalls, since proteins such as albumin frequently also function as carriers for protein fragments of biological interest [35], [36]. These technical obstacles have hitherto hampered a comprehensive analysis of the most complex proteomes by altering the outcomes with minor, albeit inevitable, loss of information [37].

View Large Image Download to PowerPoint

Fig. 1. Recently, pre-fractionation methods have been referred to as a panacea to complex samples analysis [34]. Of all the procedures, combinatorial ligand libraries appear to be the most appealing solution. Hexapeptide ligand libraries could be packed into the stationary phase of affinity chromatography columns. Complex samples, such as plasma or serum, are introduced in the column. High- and low-abundant proteins both interact with different and specific libraries and are retained in the stationary phase. However, high-abundant proteins rapidly saturate their targeted bead ligand libraries, thus they readily flow through the column. Multiple wash steps could be performed in order to clean the column from unbounded components. In the end, elution is performed. In the eluted fraction, high-abundant species result to be drastically reduced (but not absent), while low-abundant species have been now concentrated. As a result of equalization, new proteins are now perceivable through classic proteomic analyses (e.g. 1D-SDS-PAGE), enabling detection of the “hidden proteome”.

1.2. The proteomics workflow 

Proteomics analysis actually begins at the end of the sample preparation. Protein species undergo an analytic step which separates them on the basis of their biochemical/physical properties (e.g. molecular weight, isoelectric point, mass/charge ratio). This analytic phase mainly relies on gel-based approaches (mono- or bi-dimensional electrophoresis) and chromatographic methods [23], [24], [25], [26], [27], [28] (Fig. 2). Separated protein spots are then cut from the gels and trypsinized (thus cleaved into peptides) or directly chromatographically eluted to a mass spectrometer for protein/peptide identification (also known as peptide mass fingerprinting). The protein from which these peptides were derived is determined upon mass spectrometric identification by comparing the obtained sequence with theoretical mass predictions of “known” protein sequences from the database.

View Large Image Download to PowerPoint

Fig. 2. The proteomics workflow. As reported in the text, upon sample collection it is possible to perform pre-fractionation techniques in order to normalize the relative quantities of protein species within blood-derived samples. This is particularly necessary in those samples with high dynamic ranges of protein concentrations such as plasma and red blood cells. Proteins or peptides could be addressed with complementary approaches, mainly relying on gel-based techniques such as electrophoresis (first two images on the left) or chromatography. These analytical techniques allow separation of the protein/peptide species by exploiting their biochemical/physical characteristics (molecular weight, mass/charge ratio, isoelectric point, etc.). Separated species are subsequently identified with mass spectrometric tools (right column images). Each protein/peptide is characterized by a unique amino acid sequence which represents its specific molecular fingerprint. Informatic elaboration of the obtained sequences is performed through comparison against online international databases. A score is attributed to each fragment/sequence cluster individuated both experimentally and in the database. The higher is the score, the more reliable is the identification. The sensitivity and high-throughputness of this approach is extraordinary, in the order of the pico- to the femto-moles.

In detail, a mass spectrometer roughly includes an ionization source, a mass analyzer, and a detector. The ionization source produces gaseous ions from molecules in either a solution or solid phase. The mass analyzer measures the mass-to-charge ratio of these ionized molecules [21]. The most common mass analyzer (time of flight or TOF) determines the mass-to-charge ratio by measuring the time required for the ions to pass through a charged field. To further increase the resolving power of the system, tandem mass spectrometers contain two mass analyzers in a row.

Huge amounts of data are produced by mass spectrometers at the end of the analyses. Thus, it has been necessary to introduce informatic platforms for elaboration of data, in order to compare them against online databases (for example MASCOT). This is perhaps the most delicate phase of the whole proteomics analysis, along with the preliminary sample preparation steps. Data analyses and elaborations should be interpreted in the light of technical and biological variability. This is probably the main reason why, although serving its role as a powerful and highly sensitive research tool, proteomics has not hitherto found a proper collocation in routine clinical practice.

Biological complexity of proteins from blood and blood components is increased by a series of post-translational modification (PTM) events, such as phosphorylations and glycosylations, which are closely related to physiological events (e.g. the composition of sugar moieties of a protein and protein aging in vivo). PTMs exacerbate the variability of the proteome framework and nullify the efforts to individuate a standard proteomic profile of healthy (and pathological) cells.

The interactome, the record of protein–protein interactions, could be of interest in order to understand the interactions among the proteins individuated experimentally. In fact, half of the 300 proteins identified in plasma until 2002 are smaller than the 45-kDa cutoff limit for kidney filtration, thus they must exist as protein complexes not to be cleared from the bloodstream [29]. If not standardized, technical caveats behind different proteomics approaches could end up influencing the experimental outcome [38]. This is relevant when considering proteomics applications to transfusion medicine (TM).

One of the main goals of TM is to ensure safety, efficiency and effectiveness of blood components and raw materials for biopharmaceutical fractionation [39]. The three mainly transfused blood product types are erythrocyte concentrates, platelet concentrates (PCs) and fresh frozen plasma (FFP) [24]. In the last few years, TM units have started to be committed with the collection, storage, banking, manipulation and reinfusion of peripheral blood stem cells either for autologous or for homologous transplantation. Granulocytes or lymphocytes can be collected as well, the former being usually transfused into neutropenic recipients with uncontrolled infectious complications [40].

All the procedures performed at blood banks, from collection, processing, testing, production to storage and delivery of blood components, are strictly regulated by laws and/or directives issued by State or government agencies. Previous efforts in TM research, as a heritage of reductionist biology, have been so far aimed to identify single biomarkers to be adopted as diagnostic factors in ordinary clinical analysis [25]. Proteomics instead allows a comprehensive study of protein modifications, yields qualitative and quantitative information and high-throughput protein identification with unprecedented specificity and sensitivity. Therefore, proteomics potentially enables a global assessment of processing, pathogen reduction and storage methods, as well as of possible contaminants and neoantigens which may influence the immunogenic capacity of blood-derived therapeutics [28]. There still remains in TM an ambitious agenda which includes the determination of proteomic profiles to associate with healthy/pathological phenotypes, or as a consequence of the blood bank or industrial manufacturing processes [9], [14], [41].

The objective of this review is to provide a brief overview of published proteomic studies regarding TM, which mainly addressed the characterization of blood product proteomes and their modifications upon production or storage processes. A glance will be given at translational studies, in order to show the likely potential of this novel approach in the clinical endeavour.

2.1. From sickle cell anemia to storage issues: clinical proteomics of red blood cells (RBCs) 

RBCs play a pivotal role in gas transport (i.e. oxygen, carbon dioxide and nitric oxide [42]) and a minor, but not least important, role in a range of other functions, as they transfer GPI-linked proteins [43], [44] and transport of iC3b/C3b-carrying immune complexes [45].

In humans, the circulating mature RBC is the end stage of a developmental process which starts in the bone marrow, as hematopoietic stem cells differentiate to enucleated reticulocytes [46]. After extrusion of the nucleus and degradation of internal organelles and endoplasmic reticulum, reticulocytes emerge in the circulation, where they rapidly (24h) develop into mature RBCs [47]. Until the end of its life span (120±4 days), with 120 miles of travel and 1.7×105 circulatory cycles, the human RBC has successfully coped with a number of dangers, such as passages across narrow capillaries and splenic slits, periodic high turbulences and high shear stresses, along with extremely hypertonic conditions.

Erythrocyte concentrates are still the most transfused blood component worldwide [48]. Therefore, it is no wonder that they have growingly attracted a great deal of attention both in the clinical and academic settings. The absence of a nucleus and internal organelles definitely eases protein-oriented approaches to RBCs, while it disqualifies mRNA-targeting methods, such as microarrays. Nonetheless, Classic investigations have so far addressed only a handful of RBC proteins at once, whereas proteomics has recently offered the possibility to explore a panoramic view of the RBC protein complement to the genome proteome in a fast and high-throughput way (Table 1).

Table 1. Relevant RBC proteomic studies.

	Authors	Year	References
	Basic research
	Rosenblum et al.	1981	[49]
	Low et al.	2002	[50]
	Körbel et al.	2005	[54]
	Goodman et al.	2007	[51]
	Roux-Dalvai et al.	2008	[52]
	D’Alessandro et al.	2009	[80]
	Diseased versus healthy RBCs
	Jiang et al.	2003	[55]
	Florens et al.	2004	[53]
	Kakhniashvili et al.	2005	[56]
	Chou et al.	2006	[57]
	Prabakaran et al.	2007	[60]
	Storage
	Messana et al.	2000	[72]
	Annis et al.	2005	[73]
	D’Amici et al.	2007	[71]
	Bosman et al.	2008	[74]

Early analyses have purported to create a preliminary inventory of RBC proteins, at first relying on gel-based methods at the end of the 1980s [49]. If in the one hand classic approaches have yielded relevant results, on the other hand only the rapidly evolving field of proteomics can guarantee detection of hundreds of proteins in a single experiment [50]. This huge amount of data could be suitably elaborated with informatic tools in order to provide maps of protein–protein interactions, in likewise fashion to the interactome draft proposed by Goodman et al. in 2007, in which the authors drew a preliminary map of interactions between 751 RBC proteins [51]. This kind of analysis has won valuable understanding of RBC physiology, suggesting that an important fraction of the RBC proteome is devoted to respond to oxidative stress and protect proteins from folding-related issues upon oxygen-induced injuries [51].

Although proteomics has already contributed a massive amount of data, the list of RBC proteins is far from being conclusive. Revealing new entries is a challenging task because of the unbalanced proportion of RBC proteins, since hemoglobin approximately accounts for 90% of the dry weight, 98% of the cytosolic proteome. Several pre-fractionation procedures [30], [31], [32], [33], [34] have been introduced which yielded improved results and shed light on a whole concealed proteome of low-abundant protein species [52]. In 2008, Roux-Dalvai et al. revealed 1578 cytoplasmic proteins upon pre-fractionation of RBCs through combinatorial ligand libraries [52].

Each cell population has its own specific proteomic profile, since the proteome fingerprinting is strongly influenced by cellular conditions (healthy, diseased, pathological). These physiological differences have been adopted as a strategy to distinguish among subsets of the same cell population under mutated conditions. In this view, Florens et al. [53] evidenced two specific proteins which are differentially expressed on the surface of Plasmodium falciparum infected RBCs versus healthy cells, thus determining two peculiar key targets for drug effective treatments. However, whether it is not always as easy to relate single molecules with pathological/diseased conditions, proteomics approaches could deliver a panoramic view of the changes upon stimulation of pivotal receptors (for example erythropoietin receptor [54]).

Proteomics tools could be adopted to perform functional studies as well, as in the research by Jiang et al. [55]. The Authors performed a comparative analysis of healthy RBCs and RBCs from patients suffering from type-2 diabetes. Gel-based and mass spectrometric tools evidenced 27 spots being up-regulated in diseased RBCs (such as lipid raft protein flotilin), while 15 were down-regulated. The decreased expression of synthaxin, a protein mediator of target-membrane fusion, was suggestive of a misregulation of glucose transport.

Another relevant body of functional proteomics investigations is represented by the studies on patients affected by sickle cell anemia [56], [57], [58]. Indeed, it is possible both to compare sickle RBCs against healthy erythrocyte populations in order to reveal a distinctive phenotype [56], [57] and to assess the effect of ad hoc drug treatments on these cell populations [58]. In the first study, Kakhniashvili et al. [56] individuated a series of changes in lipid raft components (e.g. flotilin, stomatin) and oxygen radical scavengers (e.g. peroxiredoxins, catalases). These alterations have been suggested to play a compensatory, beneficial role against the oxidative stresses from which sickle RBCs suffer. The individuated proteins did not appear to be responsible for the clinically relevant problems of sickle cell patients, vasoocclusive crisis above all. In this respect, Goodman's group has recently applied proteomics technologies to the follow up of hydroxyurea treatment to patients affected by sickle cell anemia [58]. Hydroxyurea effectively reduces sickling in the management of homozygous sickle cell anemia, accounting for a characteristic mechanism which in part involves increased expression levels of fetal hemoglobin. The authors individuated a series of over-induced proteins upon hydroxyurea treatment, which included anti-oxidant enzymes such as catalase, thioredoxin peroxidase, biliverdin reductase and the chaperonin containing TCP1 complex (which assists the folding of RBC cytoskeletal proteins). These data could complement current knowledge on the deduced role of hydroxyurea in reducing neutrophil-mediated lipoxigenase-induced phosphatidylserine-exposure on the outer leaflet of sickle RBC membranes. This phenomenon triggers aggregation with nearby cells, which primarily determines vasocclusive crisis [59]. Although these results are not conclusive, these observations contribute to strengthen the belief about the central role of oxidative events in sickle cell anemia. Analogously, oxidative damage to RBCs is deemed to constitute (or to be related with) a peripheral component of some central nervous system disorders, such as schizophrenia [60]. In 2007 Prabakaran et al. individuated 49 spots whose expression varied between RBCs from healthy controls and schizophrenic patients. It is of note that several quenchers of reactive oxygen species (ROS) (e.g. selenium binding protein 1, thioredoxin and glutathione reductase) were found to be over-expressed in the latter group [60].

Recently, proteomics studies on RBCs have been focused on storage issues. RBC transfusion is extremely relevant for the maintenance of the oxygen transport capacity and its delivery to the tissues in anemic patients [61]. Nonetheless, the ideal storage protocol has not yet been discovered and countless attempts are being made by both blood bankers and manufacturers in this field [62], [63].

Storage of RBCs in polyvinyl chloride (PVC) bags in presence of anticoagulant and preservative solutions is performed at 1–6°C for a maximum of 42 days. At the end of the storage, haemolysis should be below 0.8%, while in vivo recovery at 24h from transfusion must be over the 75% threshold. Addition of rejuvenation solutions could diminish the macroscopic events occurring during storage, such as formation of membrane spicules resulting in the echinocyte phenotype. These solutions further improve by 2 or 3 weeks the shelf-life of stored RBCs through the reactivation of the cell metabolism upon consumption of the glucose, ATP and 2,3-DPG within the unit. Rejuvenation solutions (for example Rejuvesol) re-raise the levels of pyruvate, inosine, phosphate and adenine in order to meet the cellular demand [64]. Another accepted protocol implies liquid nitrogen or −80°C frozen storage of RBCs after glycerolization (whose dose should be inversely proportional to the freezing temperature) [65]. This protocol allows optimal storage of frozen RBCs for up to 10 years [66], although Valeri et al. have experimentally achieved a frozen storage lasting for as long as 37 years [67]. However the high costs have so far limited its broad application in clinical practice [65]. The need for a long-lasting storage of RBCs meeting standard acceptability criteria on the one hand and the effectiveness of a low-cost procedure on the other hand will hopefully find a synthesis in the anaerobic storage protocol proposed by Yoshida and his Group [68], [69], [70] and proteomically investigated by Zolla's group [71].

So far, the quality of stored RBCs has been tested with routinely clinical tools and a few parameters have been addressed: in vivo recovery after 24h from reinfusion, which is monitored by means of 51C isotope; pO2, from which sO2 could be derived, that is measured by specific oxygen sensors; haemolysis values, glucose, DPG and ATP levels as well as pH, by means of standard biochemical approaches. Proteomics has been recently proven to be a valuable tool in quality control analysis of stored RBC units, although this is actually only limited to the academic research, which will be hopefully translated into clinical practice in the future. Several proteomics investigations have been carried-out focusing on RBCs under different storage conditions [71], [72], [73], [74]. In 2000, Messana et al. [72] found that oxygen-dependent metabolic modulation resulted to be progressively altered during storage and even addition of rejuvenation solutions at day 21 did not contribute to restore it. The authors indicated alterations of the anion exchanger band 3 and the metabolic enzyme glyceraldehyde 3-phosphate dehydrogenase as crucial factors in mediating RBC storage lesions. In 2005, Annis et al. performed an analysis of membrane protein composition of CPD–saline–adenine–glucose–mannitol (SAGM)-stored RBCs with gel-based and mass spectrometric approaches at days 1, 14, 35 and 42 of storage [73]. Much lower amounts of proteins were present in the supernatants of leukofiltered RBCs and, although the majority of identified proteins were common to both types of RBC concentrates, transthyretin (a transport/binding protein), Igk-light chain (Igk), serum amyloid P (SAP), and connective tissue activating peptide III (CTAP-III) accumulated predominantly in non-leukoreduced RBCs, whereas cytosolic enzymes, such as carbonic anhydrase I and thioredoxin peroxidase B were found to accumulate in leukofiltered RBCs. The unexpected increase of serum proteins, such as transthyretin, Igk, and SAP, can be explained by the fact that they are absorbed on the cell surface, and then released during storage [24]. CTAP-III is a peptide cleaved from a PLT basic protein and promotes neutrophil adhesion [75]. These results underline the importance of leukoreduction in abrogating the pro-inflammatory response elicited by supernatants from stored RBCs [75], since the presence of white blood cells in the storage unit could represent a burden in increasing the number and strength of RBC adhesion to vascular endothelium [77]. The accumulation of cytosolic proteins in the supernatant of leukofiltered RBCs is instead explained with storage-related haemolysis [24].

A recent article by Zolla's group examined the changes of RBC cytoskeleton during storage of SAGM-preserved non-leukodepleted RBC units either under anaerobic and aerobic conditions [71]. Leukocyte-reduction was not performed, in order to include any contribution of white blood cells to proteolytic cleavage and ROS production. The authors adopted gel-based tools for the detection of RBC membrane changes over storage, either under atmospheric oxygen or helium, in the presence or absence of protease inhibitors. The etiology of lesions in RBC membranes involves both ROS and proteolytic enzyme activity [68], [69], [70]. A gradual increase in the number of protein spots (due to fragmentation and aggregation phenomena) was observed during the first 14 days of storage, followed by a decline at 42 days. In the presence of oxygen, many new protein spots were generated as a result of cytoskeleton protein attack by ROS, whereas only a small number of changes were related to proteolytic cleavage, which seemed to play a minor role in storage lesions in comparison to protein oxidization. During the first 7 days of storage, oxidative damage was observed prevalently in band 4.2, to a minor extent in bands 4.1 and 3, and in spectrin. All those factors might contribute to exacerbate the formation of neoantigens in the blood units, which might be related to a series of untoward effects retrospectively observed in some categories of recipients (e.g. intensive care unit or traumatized patients) [76], [77]. Protein degradation was greatly reduced in the absence of oxygen, when blood was stored under helium. In agreement with Yoshida and his group [68], [69], [70], this study confirmed that any action to improve storage conditions should be carried out in the first weeks in order to prevent damages of the membrane–cytoskeleton network. To this end, oxygen removal is a more effective way of limiting RBC storage lesions than any chemical addition.

From a recent study by Bosman et al. it emerged that storage reduced the membrane-protein variability and increased the number of proteins individuated in membrane-shed microvesicles [74]. A total of 257 proteins were identified with gel-based, chromatographic and mass spectrometric approaches. It resulted that storage reduced the membrane protein variability (less band 3, small G-proteins, chaperones and components of the proteasome were observed) while it increased the total number of microvesicle-isolated proteins (especially as it regards Hb, band 3, CD47, complement proteins and metabolic enzymes). This study was focused on the global changes occurring at the membrane level during storage, as they were a direct result of disturbance and/or acceleration of physiologic processes such as cellular aging, including vesicle formation [78], [79]. Bosman and co-workers underlined the fact that early interventions should be targeted to prevent storage lesions in the first 2 or 3 weeks of storage, in agreement with Zolla's group proteomics observations [71].

At present, a proteomic analysis of frozen-stored RBCs is still missing. The elevated costs for the maintenance of frozen RBC units could become a minor issue whether proteomic analysis will assess their improved quality in respect to hypothermic stored erythrocyte concentrates.

Proteomics investigations on RBCs have so far yielded the identification of a total of 1989 proteins, as it emerged from a recent bioinformatic overview of data from literature [80]. Nonetheless, network and pathway analyses through innovative softwares seem to suggest that the proteome of RBCs will be further expanded in the near future [80].

2.2. Platelet (PLT) proteomics: collection methods, storage lesions, activation and releasates 

Human PLTs are the smallest formed elements of the blood; they are enucleated, disc-shaped, membrane-encapsulated cell fragments that are formed and released into the bloodstream primarily by bone marrow megakaryocytes [81]. In whole blood, PLT concentrations range between 150 and 400×109 cells per liter, with an average lifespan from 7 to 10 days, delimited by apoptosis [82]. One third of the total PLT pool is sequestered by the spleen. PLTs maturate from megakaryocytes and could be roughly listed as follows: a separate, membrane-delimited dense tubular system; a cytoskeletal network; a peripheral band of microtubules; specialized organelles including α- and dense granules, lysosomes, microperoxisomes, and mitochondria [83], [84], [85]. Being enucleated, PLTs do not host a proper genome and inherit meager amounts of mRNA from their megakaryocyte ancestors [86], [87], [88]; this prevents them to be an eligible target for transcriptomic analysis [14]. PLTs exert a key function in the preservation of vascular integrity and maintenance of hemostasis [89], [90].

From 1979, early approaches to PLT proteomics relied on gel-based techniques [91], [92] (Table 2). It was only in the last few years that the PLT proteome has been directly investigated. This may be due to the wide abundance range of proteins in PLTs, in which cytoskeletal proteins (mainly actin) represent the greatest portion [14]. Removal of high-abundance proteins has been attempted in PLTs, either through gel-based preliminary fractionation [93] or combinatorial ligand libraries [94].

Table 2. Relevant PLT proteomic studies.

	Authors	Year	References
	Basic research and activation
	Clemetson et al.	1979	[91]
	Gravel et al.	1995	[92]
	Immler et al.	1998	[98]
	Marcus et al.	2000	[99]
	Maguire et al.	2002	[101]
	O’Neill et al.	2002	[95]
	Marcus et al.	2003	[100]
	Garcia et al.	2004	[96]
	Garcia et al.	2004	[102]
	Claeys et al.	2005	[97]
	Moebius et al.	2005	[93]
	Lewandrowski et al.	2006	[104]
	García et al.	2006	[103]
	Guerrier et al.	2007	[94]
	Storage and releasate
	Snyder et al.	1987	[119]
	Coppinger et al.	2004	[105]
	Glenister et al.	2007	[121]
	Thiele et al.	2007	[120]
	Maynard et al.	2007	[123]
	Ryu et al.	2008	[125]
	Thon et al.	2008	[126]
	Assessment of donor suitability
	Sacristan et al.	2008	[128]
	Springer et al.	2009	[127]

In 2002 O’Neill and his group performed an analysis of the PLT proteome by means of gel-based separation protocols and mass spectrometry identification [95]. The authors individuated 284 proteins (out of 2300 spots).

Garcia et al. expanded the PLT proteome by detecting overall 760 proteins [96]. Latest surveys have been focused on the hydrophobic fractions and on membrane proteins [97]. Blue native is a useful gel-based approach which is indicated when studying hydrophobic membrane proteins. By means of blue native separation and mass spectrometric identification the authors isolated 63 proteins out of 58 spots, most of which were membrane-bound proteins [97].

Hexapeptide baits were adopted to capture and concentrate low-abundant proteins while cutting the concentrations of the most abundant ones [94], bringing unambiguous identification of 435 proteins. Out of those, 180 turned out to be precedently undescribed proteins which were added to the already rich “cartography” (more than 1100 of unique entries) of the PLT proteome.

In parallel to these basic investigations, in 1998 a functional study was aimed at detecting protein phosphorylation levels (phosphoproteome) of thrombin-activated PLTs, which were preemptively labeled with 32P [98]. Protein phosphorylation is a key regulatory event in a huge series of biological processes, from cell cycle regulation to proliferation, metabolism regulation or activation. Determining the pivotal targets of activation events could ease the engineering of specific drugs in perspective. In this respect, Marcus et al. performed two closely related investigations in 2000 [99] and 2003 [100]. Phosphorylated tyrosines were either identified by means of specific antibodies or with autoradiography of 32P-labeled proteins. In a similar study, Maguire et al. confirmed the relevance of phosphorylation events in thrombin-activated PLTs in 2002 [101]. pCas (130kDa), FAK (125kDa), PI(3)k (85kDa) and src (85kDa) were among the key targets of these PTM events [101].

Garcia's group extended this concept not only to thrombin-receptor activated PLTs [102] but also to collagen-related peptide-activated ones [103]. The authors individuated 62 [102] and 96 [103] differentially phosphorylated proteins in 2004 and in 2006, respectively. The adapter tyrosine kinase 2 (DOK-2) and the regulator of G-protein signaling (RGS), as well as glycoprotein VI (GPVI)-mediated signaling pathways were involved in PTM processes triggered by thrombin-receptor mediated activation [102], while the Grb2-platelet membrane adapter and G6f proteins were targeted by kinases (phosphorylating proteins) upon collagen-related peptide activation [103]. G6f is known to couple to the Ras-mitogen-activated protein kinase pathway in the immune system. In 2006, Lewandrowski et al. suggested that G6f could be a relevant target for glycosylations as well [104]. The Authors recognized 70 sites for glycosylation in 41 different proteins. Since differently glycosylated isoforms of the same proteins are likely to play a different function within the cell, this study conveyed new valuable understanding of the molecular mechanisms regulating PLT activation by means of PTM switches [104] and thus offered the potential to design targeted drugs to tackle these phenomena.

PLT activation does not only induce intra-cellular molecular cascades, but it also provokes alterations in the contents of releasates, which are likely related to a series of untoward effects in the reinfused recipients upon transfusion of the PLT concentrates. Coppinger et al. [105] distinguished over 300 proteins from the secretome of thrombin-activated PLTs. Many of the proteins identified were not previously attributed to PLTs, including secretogranin III, a potential monocyte chemoattractant precursor; cyclophilin A, a vascular smooth muscle cell growth factor; calumenin, an inhibitor of the vitamin K epoxide reductase–warfarin interaction. These proteins had already been related to atherosclerotic lesion, suggesting that PLTs may contribute to the etiology of this disease other than to its thrombotic complications [105].

It 1960s PLTs became a standard treatment for bleeding thrombocytopenic patients with bone marrow failure [106], [107]. PLTs are collected through apheresis from a single donor or from pooled buffy coats from 4 to 6 donors [26], [27], [28] and stored at 22°C under continuous agitation for optimal gas diffusion through gas-permeable bags. Currently PLTs are stored for up to 5 days to prevent bacterial contamination [108]. Over the past 10 years, a number of pathogen reduction methods have been reported, which enable the reduction of viruses, bacteria, and protozoa that are present in cellular blood products by means of photodynamic or photochemical methods [109], [110], [111], [112], [113], [114], [115]. Even in absence of contaminating agents, storage itself causes a variety of changes in PLT function and morphology, the so-called PLT storage lesions [116], [117]. At present, the quality of PLT concentrates is primarily determined in vitro by selective methods, such as swirling assessment, pH determination, flow cytometry or aggregometry, which can provide only limited information regarding PLT function or morphology [118]. Proteomics could potentially become a powerful tool for quality control analysis of PLTs during storage in clinical practice by evidencing the molecular alterations which take place within stored units, as it has been recently suggested by newly achieved academic advancements.

Since 1987, electrophoretic methods have been used to monitor alterations during the storage of PLT concentrates [119]. When prolonging storage from 5 to 7 days, Snyder et al. found out that 30 spots were differentially expressed and 2 actin fragments significantly accumulated over time [119].

Proteomics methods could be suitable to evaluate proteins which accumulate in the supernatants during storage [120], [121], [122] and are thus related to a series of pro-inflammatory events which are triggered upon long-stored PLT transfusion. In the first study [121], a series of cytokines (commonly implicated in allergic responses), such as BDNF, CCL5, PDGF as well as clusterin (a complement-mediated lysis inhibitor), TLT-1 (having a role in PLT adhesion) and ILK (aggregation and adhesion to damaged endothelium) were found to increase in the supernatant after 7 days of storage, suggestive of progressive leakage from or degradation of PLTs during storage. These findings are relevant in light of the current discussion around the safety and efficacy of extending the shelf-life of PCs to 7 days instead of the present 5-day expiry. Concerns have so far only addressed the increased risk of bacterial growth, proportional to the prolongation of the storage. Longer-stored PLTs actually display a higher pro-inflammatory capacity, even when leukoreduced, which may correlate to the persistence of adverse transfusion reactions (such as febrile non-haemolytic reactions) thus nullifying the efforts for a quantitatively improved storage. In other terms, it is not only a matter of the increased contamination risk, but also of the compromised quality of the longer-stored PC.

Proteomics could allow to test the quality of PLTs during storage, but also to assess the advantages and the drawbacks of the one collection method over the other. To this end, the supernatants of apheresis and buffy coat platelet concentrates were monitored by Wurtz et al. [122] with chromatography and mass spectrometry. PLT factor 4 (PF4) and β-thromboglobulin (β-TG) concentrations were used as a parameter for PLT activation. These proteins are contained in the α-granules [123] which are secreted upon activation. Their concentration (after 24h) in the supernatants of buffy coat PCs was lower in comparison to PLT collected through apheresis, thus suggesting a minor activation of the former concentrates.

Further data from Glenister et al. confirmed that during the first 24h after donation, the apheresis PLT concentrates showed major deviations in their proteome in comparison to buffy coat ones: owing to the preparation procedures, buffy coat PCs might have had a longer time to recover from preparation stress [121]. The methods used by the authors could be useful for the evaluation of different PLT collection, production and storage protocols as well as the performances of different cell separators and of the recently introduced streamline and automatic buffy-coat processor devices. Thiele et al. [120] found that septin 2, gelsolin, and β-actin (respectively apoptotic and activation markers) were altered at the end of the storage. These observations reinforced the relationship between the short shelf-life of PLTs and apoptosis [81], [124].

Leukofiltration is a relevant step of platelet concentrate preparation in order to avoid side effects after reinfusion to the recipient. To this end, a proteomic screening of PLTs stored for 24 and 120h showed that the total number of protein spots increased with storage and decreased after filtration [125]. In particular, macrophage inflammatory protein-2 alpha, megakaryocyte colony stimulating factor and interleukin-22 changed with storage and leukoreduction. In the light of all the mentioned results, storage seems likely to cause both activation and apoptotic events in PLTs.

These observations suggest that protocols contrasting the apoptotic cascades might eventually prolong PLT shelf-life and enhance their viability during storage, as Li had already suggested in 2000 [82].

Recently, Devine's group adopted complementary proteomics approaches to monitor PLT storage lesions from day 1 to 7 of storage [126]. These approaches were referred to as protein-centric or peptide-centric, since they addressed the single molecules or protein aggregates/complexes, respectively. The authors evidenced variations of 503 protein expressions during the whole period, although membrane proteins were only revealed by means of peptide-centric approaches. These results prompt a series of considerations about the use of proteomics in clinical practice, whereas the choice of the method to be adopted should be guided by a balance between the most suitable technique for the question being posed and the costs for the required facilities. Most comprehensive analyses should be carried out integrating both the peptide- and the protein-centric approaches.

Recent papers on PLT storage proteomics introduced a revolutionary approach, since they show how it could be potentially possible to address the suitability of the donor (or donor category) through a thorough assessment of the quality of the final product. In this view, Springer et al. recently compared PLTs from healthy donors and donors affected by type-2 diabetes [127]. Out of a total of 844 individuated proteins, 122 were either up- or down-regulated in type-2 diabetics relative to non-diabetic controls. Quantitative analyses evidenced fluctuations in the abundance of 117 proteins, mainly diminishing their expression during the 5-day storage period [127]. This is the first study shifting the focus of attention from the final product to the provider (donor). Indeed, a major consideration arises after this study concerning the potential use of PLTs from diabetic donors, which behave like long-stored PLTs even when fresh. These PLTs rapidly age and activate during storage, resulting in a compromised hemostatic function [127]. Taken together, these findings make it desirable to further investigate the suitability of diabetics as PLT donors. This approach represents a sort of Copernican revolution in TM, whereas proteomic data are not only collected to delineate the proteome profile but also for the interpretation of functional aspects either regarding the end product, or most likely, directly addressing to the suitability of its provider. Other studies, currently in progress, share the same kind of approach. Preliminary studies suggest that PLT proteome profiles differ between male and female donors (Devine Dana, personal communication, 11th European Haemovigilance Seminar, 26th February 2009, Rome, Italy; Peter Schubert, personal communication, Blood and Proteomics in Viterbo, 13th October 2009, Viterbo, Italy).

Analogous studies seem to question the suitability of PLT donors which suffer from moderate hypertension [128]. Many of the clinical events associated with hypertension are related to thrombotic processes including stroke [128]. In this regard, there is evidence showing that, in hypertensive patients, the sensitivity of PLTs to become aggregatory agents is increased, an effect that is accompanied by basal PLT pre-activation. Sacristan and colleagues have recently investigated the effects of the administration of Olmesartan medoxomil (an angiotensin II receptor blocker) on the PLT proteome of patients affected by moderate hypertension. Gel electrophoresis maps evidenced variability among healthy controls, hypertensive patients (either treated or untreated), suggestive of an actual activation of PLTs from the latter groups. Besides, these observations raise concerns about the characteristics of PLTs from hypertensive patients, even when under treatment with ad hoc drugs. In fact these patients might routinely be accepted as PLT donors, while the quality of their PLTs is questionable, in like fashion to Springer's findings on diabetic patients [127].

In 2009 Garritsen et al. [129] widened the list of the possible applications of proteomics in PLT-TM. The authors adopted a specific mass spectrometric technology for efficient genotyping of the human PLT-specific antigens (HPAs), whose clinical relevance is demonstrated by several immune-mediated PLT disorders [130], such as alloimmune thrombocytopenic syndrome in neonates which passively receive maternal antibodies against fetal HPAs [131]. In transfusion medicine, mismatches between donor and recipient for HPAs can cause refractoriness to PLT transfusions and post-transfusion purpura [132]. Moreover, HPA genotyping could improve the success of PLT transfusion [133]. Although RBC, PLT and white blood cell genotyping is a hot issue in TM and several genomic approaches have been successfully proposed [134], proteomics could rapidly grow as a serious competitor due to its complete automation, high-throughputness, extreme rapidity of the multiplexed analysis, cost-effectiveness, very high sensitivity and specificity. Indeed, although the initial purchase of mass spectrometric equipment is at this moment rather costly, the running costs for assays like the one presented by Wu and Csako can be very low (3.5 cents per genotype) [134].

2.3. Plasma proteome: mare magnum of complexity 

Human blood plasma is an exceptional proteome, also containing other tissue proteomes as a subset [29], intact, as well as partially degraded proteins or protein fragments [14]. Plasma proteome is a dynamic entity since it is influenced by stress, sleep, sport training, or meals and, in women, by pregnancy [14]. A series of issues should be faced when performing roteomics investigations of plasma, such as the large proportion of albumin (55%, in the order of 30g/L), the wide dynamic range in abundance of other proteins, and the tremendous heterogeneity of its predominant glycoproteins [29]. A handful of proteins make up the greatest portion of plasma proteins, among which immunoglobulins (Igs), fibrinogen, transferrin, haptoglobin and lipoproteins are the main constituents along with albumin. Therefore, the removal of high-abundant proteins is particularly essential to reveal low-abundant species in plasma. Likewise, clotting induction causes the disappearance of several proteins involved in the coagulation process such as prothrombin (Factor II), which is absent in serum, while it enables detection of many otherwise hidden proteins from the serum proteome [135]. Plasma proteins have been studied since before we knew that gene existed, yet still the completion of the plasma proteome is far from being at hand (Table 3). Plasma is obtained from whole blood fractionation or by apheresis. It is produced by blood banks as fresh frozen plasma (FFP) for the treatment of bleeding due to coagulation disorders, or as raw material for biopharmaceutical fractionation, to manufacture medicinal products [135]. The history of plasma proteomics is deeply rooted in the history of proteomic itself [26]. Since Tiselius found that serum could be fractionated into multiple components on the basis of electrophoretic mobility [136], a long list of papers have been published dealing with plasma proteins and technical approaches to investigate them. However, it was only in 1977 that Anderson and Anderson [137] performed the first modern screening of the plasma proteome through gel-based techniques, as proposed by O’Farrell in 1975 [138]. The same approach could be helpful to reveal polymorphisms in the proteome pattern of serum specimens from a normal population [139] or from different populations [140] or to evidence the glycosylation pattern by means of lectin blotting revealed with chemiluminescence [141]. An example of the latter case is the gel-based analysis of plasma-protein glycosylation patterns in patients with chronic alcohol abuse [142]. Of note, alcoholism apparently induced abormal transferrin, haptoglobin β and alpha(1)-antitrypsin isoforms, devoid of a variable number of entire N-glycan moieties [142]. These observations have obvious clinical relevance, since they could be early signals of this kind of dependence.

Table 3. Relevant plasma/serum proteomic studies.

	Authors	Year	References
	Basic research and applied quest for suitable biomarkers
	Tiselius	1937	[136]
	Anderson and Anderson	1977	[137]
	Tracy et al.	1982	[139]
	Harrison et al.	1991	[140]
	Gravel et al.	1994	[141]
	Henry et al.	1999	[142]
	Richter et al.	1999	[143]
	Anderson and Anderson	2002	[29]
	Omenn et al.	2005	[144]
	Li et al.	2005	[145]
	Janzi et al.	2005	[159]
	Smalley et al.	2007	[146]
	Qian et al.	2008	[154]
	Schenk et al.	2008	[155]
	Zheng et al.	2009	[156]
	Omenn et al.	2009	[158]
	Muthasamy et al.	2008	[157]
	Pathogen reduction
	Tissot et al.	1994	[148]
	Crettaz et al.	2004	[152]
	Steil et al.	2008	[153]
	Plasma derivatives
	Demelbauer et al.	2005	[200]
	Plematl et al.	2005	[201]
	Brigulla et al.	2006	[197]
	Yang et al.	2009	[203]

Mass spectrometry represented an actual revolution in the field of plasma proteomics as well [143]. During the first 25 years of proteome research activity on plasma, the plasma protein mosaic was greatly expanded by one order of magnitude [29]. As reviewed by Anderson and Anderson [29], in 2002 the plasma protein virtual database enlisted approximately 300 proteins. Only 3 years later, the Human Proteome Organization (HUPO) in merit of the Plasma Proteome Project (PPP) regrouped data from 35 collaborating laboratories which merged into a total of 3020 proteins [144]. The burgeoning amount of complementary approaches and the refinement of preliminary sample treatments, such as high-abundant protein depletion, definitely contributed to reach this goal. Strikingly, another study from a single laboratory individuated 560 proteins, out of which only 257 merged with the HUPO-PPP protein list [145]. This introduction suggests that new plasma protein discovery constantly keeps the pace with the technical advancements and a comprehensive tabulate of proteins is far from being at hand.

In 2007, Smalley et al. examined the contents of plasma and PLT-derived microparticles [146]. Quantitative mass spectrometric analysis revealed 21 plasma-microparticle peculiar proteins, among which there were proteins associated with apoptosis (CD5-like antigen, galectin 3 binding protein, several complement components), iron transport (transferrin, transferrin receptor, haptoglobin), immune response (complement components, immunoglobulin J and kappa chains), and the coagulation process (protein S, coagulation Factor VIII) [146].

As for other blood components, proteomics application could be translated from basic research to a thorough quality assessment of all the steps in the FFP production and pathogen reduction processes. A variety of procedures are applied for pathogen reduction, but it should be emphasized that none of the currently applied methods inactivates all types of pathogens, and all have some effect on plasma quality when compared to FFP [147]. Pooled solvent/detergent (S/D)-plasma is the best-documented clinical product, followed by methylene blue (MB) light treated-plasma. Recently, psolaren light treated-plasma has been introduced (CE-marked product in Europe), while riboflavin light treated-plasma is still under development. The effect of pathogen reduction treatments on the integrity of plasma proteins has been addressed by several studies. In 1994 Tissot et al. sought to verify whether MB in combination with visible light induced protein alterations [148]. The authors adopted high-resolution gel-based methods to test both untreated and photochemically treated apheresis FFP [148]. This study revealed no change in the protein pattern before and after the treatment. Nevertheless, the problems related to MB potential genotoxicity have to be overcome [149]: removal of MB by filtration has been proposed as one of the effective efforts to efficiently remove the dye after photo-treatment [150], [151]. In 2004 Crettaz et al. examined the effect of MB treatment upon light exposure, followed by dye removal with different MB concentrations [152]. Modifications were noticed in the γ-chain of fibrinogen. Concomitantly with functional alterations of this protein, the authors observed a relatively large increase of one of the isoforms of transthyretin, and an enlargement of the spot of apolipoprotein A-I. Removal of MB by filtration did not cause additional significant protein alterations. The effect of over-treatment of plasma with very high concentrations of MB (50μM) in association with prolonged light exposure (3h) showed severe damage to most of the plasma proteins. By contrast, the standardized S/D inactivation method was, apparently, not associated with significant modifications of the gel-electrophoretic profile of plasma proteins [14], [26]. Few investigators utilized proteomics strategies to evaluate the effect of pathogen reduction on blood plasma preparations, but these findings may be important for the quality control of the various plasma preparations available on the market, and may reveal protein modifications and/or otherwise unexpected neoantigens. In 2008 Steil's group analyzed freeze-dried S/D-treated human plasma [153]. Samples from 1000 patients were pooled together, lyophilized and analyzed after 24 months of storage. Gel-based approaches evidenced that lyophilization of human plasma neither altered its protein composition nor impaired its clotting capacity. The only alterations occurred during the first steps of S/D inactivation and reflected known effects of the S/D treatment. These features of the product make it an attractive option in situations where cold chains cannot be warranted and help to provide plasma with reduced delay in emergency situations.

The search for new plasma proteins is far from being concluded. Many authors have still been contributing to the cause in the last year [154], [155], [156]. Currently, a total of 7518 proteins and isoforms, 5234 PTMs and 3778 unique genes are tabulated in the plasma proteome database available online [157]. The database is a unique list of unambiguously identified proteins with a peculiar focus on the site of expression and proteolytic cleavage [157], since PTMs and quantification methods are the declared prospective targets of the HUPO-PPP [158].

Paradoxically, even if the plasma proteome has been dramatically expanded over the last decade, in the United States the rate of introduction of protein tests approved by the Food and Drug Administration has declined to less than one new protein diagnostic marker per year [29]. The lack of a holistic view has hitherto hindered the path of plasma proteomics, thus impeding translation of basic research achievements into clinical practice. Hence, in this phase we need to seek shortcuts through proteomics, analogous to the path that shotgun sequencing blazed through the genome, which are not guaranteed to exist but will prove out fundamental to depict the final portrait of healthy plasma that we need to establish the basis for further clinical applications in perspective.

When the quest for suitable biomarkers will proficuously end, daily clinical practice will be hopefully taking advantage of innovative methodologies which allow high-throughput standardized screening of massive protein datasets instead of looking for a single biomarker: the protein chips [159]. Although these tools are efficient and effective, it is still too expensive to start a laboratory endowed with such a peculiar instrumentation, even if routine costs are definitely cut once the laboratory has been equipped [159]. A series of papers have been recently published dealing with non-properly transfusion medicine-related topics, which exploit plasma/serum analysis to detect early markers of breast [160], [161], pancreatic [162], hepatocellular carcinoma [163], prostate [164] and bladder [165] cancers, just to give a few examples of 2009 only reports.

2.4. Proteomics and white blood cells: paving the way for a broader clinical application 

Although mature white blood cells are not a widely used therapeutic product, a renewed interest in granulocyte transfusion therapy has been generated by the observation that large doses of granulocytes can be obtained from donors who have been stimulated with specific recombinant growth factors [166], [167].

2.4.1. Granulocytes 

Granulocyte concentrates are prepared from the blood of healthy donors by apheresis and are therefore also known as granulocyte apheresis concentrate. To ensure sufficient granulocyte content, donors are pre-treated with corticosteroids and/or recombinant growth factors for granulocytes (granulocyte-colony stimulating factor, G-CSF). This pre-treatment with G-CSF significantly increases granulocyte yield and prolongs granulocyte survival time [168], [169].

Granulocyte concentrates are used for the treatment of severe bacterial and fungal infections in neutropenic patients undergoing hematopoietic stem cell (HSC) transplantation and dose-intensive chemotherapy for malignant diseases [27]. Neutrophils are the main class of granulocytes and constitute the largest white blood cell population in human peripheral blood. Morris et al. have recently highlighted a timeline of the major steps given throughout the years towards a deepened knowledge about the biology of the neutrophil [170] (Table 4). Nevertheless, there is still a limited number of papers addressing the neutrophil proteome by different standpoints: calcium-dependent secretion [171], tyrosyl-radical attack targets as a hallmark of neutrophil-mediated injury at the inflammatory locus [172], the content of the granules [173], [174], the proteome of resting neutrophils [175] and the effects of GM-CSF stimulation [176].

Table 4. Relevant white cell proteomic studies.

	Authors	Year	References
	Neutrophils
	Boussac and Garin	2000	[171]
	Avram et al.	2004	[172]
	Lominadze et al.	2005	[173]
	Feuk-Lagersedt et al.	2007	[174]

2.5. Therapeutic stem cells: proteomics for transplantation purposes 

In recent years, HSC transplantations have become to be considered the eligible treatment for an increased number of malignancies and hematologic disorders as well as of several inherited diseases [177] after chemotherapy and/or radiotherapy [178], [179]. HSCs have been so far mainly derived from bone marrow, though in the last 5 years the registered trend has moved towards an increase in the harvest of HSCs from alternative sources, such as peripheral blood and umbilical cord blood [180]. More than 25 years have passed since the demonstration that HSCs can be cryopreserved [181]. Proteomic studies have been so far mainly performed on hematopoietic progenitor-enriched fractions from umbilical cord blood sources and have addressed a specific population, namely CD34+ [182], [183], [184], [185], [186], [187] (Table 5), as it has been recently reviewed [188]. CD34 is a transmembrane mucin-like protein is used as a marker for immature hematopoietic stem and progenitor cell-enriched populations, whose plasticity and proliferative capacity play a pivotal role in marrow reconstitution after the treatments [189].

Table 5. Relevant HSCa proteomic studies.

	Authors	Year	References
	Zenzmaier et al.	2003	[182]
	Wei et al.	2003	[184]
	Tao et al.	2004	[185]
	Zenzmaier et al.	2005	[183]
	Liu et al.	2006	[186]
	Clinical application
	Kaiser et al.	2004	[190]
	Wang et al.	2005	[192]
	Weissinger et al.	2007	[191]
	Imanguli et al.	2007	[193]
	Colquhoun et al.	2009	[194]

a CD34+ stem and progenitor cells from umbilical cord blood.

The most comprehensive study on human cord blood-derived CD34+ cells to date has been performed by Liu and colleagues in 2006 [186]. Their study addressed both the proteome and the transcriptome of CD34+ cell fractions. Notably, the authors identified a central core of vesicular transporter and heat shock proteins, along with a series of nerve, gonad, and eye-associated proteins, thus providing supporting evidence about a broader range of protein expression in immature cells. It could be concluded that protein expression variability seems to decrease during differentiation, while expressed proteins become more specific in the mature differentiated cell.

Recent proteomic approaches have sought to indirectly investigate the outcomes of HSC transplantations, in particular from umbilical cord blood sources [190], [191], [192], [193]. Several groups have succeeded in discriminating healthy patients and patients with graft-versus-host disease upon HSC transplantation, through a proteomic analysis of several biological fluids of the recipients, such as urine [190], [191], saliva [192] and plasma [193].

Furthermore, Colquhoun et al. have recently assessed that smoke affects the quality of umbilical cord blood sera [194] and, obviously, the health of the newborn. These translational studies have actually shed light on the widespread potential of proteomics tools and hopefully paved the way for valuable routine applications.

2.6. Other plasma derivative products 

Protein products fractionated from human plasma are an essential class of therapeutics often used as the only available option, in the prevention, management, and treatment of life-threatening conditions in consequence of a trauma, congenital deficiencies, immunologic disorders, or infections. Plasma for fractionation is obtained by plasmapheresis or from whole blood as a by-product of RBC production (recovered plasma). The plasma fractionation technology still heavily relies on the cold ethanol fractionation process, but has been improved by the introduction of modern chromatographic purification methods, and efficient viral inactivation and removal treatments, ensuring quality and safety to a large portfolio of fractionated plasma products. Many authors have reviewed the history of plasma protein therapies [195]. The rapidly evolving scenario forced the attention of the consolidated plasma industries to shift from albumin, for the treatment of battlefield injuries in Second World War, to Factor VIII for haemophilic patients [195]. Currently, the leading plasma derivative product is the intravenous immunoglobulin G, which has replaced Factor VIII and albumin in this role [196]. Recently, the analysis of prothrombin complex concentrates (PCCs) was used as a model to evaluate the sensitivity of proteomics technologies in blood-derived therapeutic products, beyond that of standard quality control [197]. PCCs contain the procoagulant zymogens of the prothrombin complex, namely Factor II, Factor VII, Factor IX, and Factor X, as well as the inhibitory factors protein C and protein S. They are currently the eligible treatment in patients with hereditary or acquired deficiencies of vitamin K-dependent clotting factors [198], and are standardized by their Factor IX activity [199]. In 2006 Brigulla et al. analyzed three PCCs (two batches each) for differences in protein content by functional proteomics assays (Table 3). The results were compared to those of a pool of 72 normal plasma samples [197]. The authors found major differences between products of different manufactures, whereas no significant batch-to-batch variability was seen in the same product. Besides the labeled clotting factors, 41 additional proteins were identified, such as fibrinogen, complement factors, and several apolipoproteins. Many proteins were present in multiple spots (Factor II, Factor X, protein C, vitronectin), indicating a high degree of PTMs. Interestingly, protein expression in the gel-electrophoretic pattern did not correlate with the activities of clotting factors, suggesting a loss of biologic function during the manufacturing process. In comparison with untreated pooled plasma, PCCs displayed several low-molecular-weight variants of proteins that likely constitute potential degradation products.

Over the last decade, Josic's group has covered a wide spectrum of basic research approaches with proteomics tools, such as for example the changes in the pattern of glycosylation of human plasma-derived antithrombin, a post-translational modification which intuitively modulates its biological activity/half life [200], [201]. Besides, the authors focused on a series of methods for the optimization of plasma protein separation [202] and detection of low-abundant plasma-derived therapeutic proteins (such as vitamin K-dependent clotting factors and inhibitors) and potentially harmful contaminants [203]. However, spreading and application of the concepts gained by means of these mere proteomic approaches have not hitherto found a broad resonance. These observations could constitute a preliminary body of experimental experiences with a great potential in quality control and possible improvement of the industrial processes of plasma fractionation.

α1-Antitrypsin (A1AT) deficiency is an autosomal recessive disorder characterized by quantitative and qualitative abnormalities of α1-proteinase inhibitor [204], which is a potent inhibitor of serine protease. In particular, α1-proteinase inhibitor especially inhibits neutrophil elastase, which is responsible for the degradation of connective tissue in the lung. Therefore, α1-proteinase inhibitor deficiency increases the risk of emphysema pneumothorax and chronic obstructive pulmonary disease [205], [206]. Recent proteomic studies assessed a diminution of A1AT in patients with idiopathic pulmonary arterial hypertension [207].

A1AT augmentation therapy is indicated for the treatment of patients with lung emphysema secondary to congenital A1AT deficiency [206]. Since the mid-1980s, A1PI deficiency has been treated with specific plasma-derived concentrates [196]. At present, there are several licensed A1AT plasma-derived concentrates. Three commercially available products were recently submitted to biochemical analysis of the protein backbone as well as to mass spectrometric analysis. A glycoproteomic characterization of the A1PI protein was performed intending to identify and characterize potential alterations, caused by the exposure to different physicochemical conditions and various enzymes, during the industrial manufacturing process [197]. The study showed that these commercially available A1PI products differ from A1PI directly analyzed on plasma, but the structural variations observed between products do not have a substantial biologic role, because all the three concentrates have similar half-lives, and specific neutralising activities. Currently, according to the Food and Drug Administration (USA), no data suggest that these differences might affect the functional activity and immunogenicity of A1PI concentrates.

Data from these studies indicate that proteomics approaches could become useful tools to assess the impact of industrial processes of plasma fractionation on the integrity of blood-derived therapeutics [197]. Comparative analyses between recombinant and plasma derivative products will be expanded in the upcoming future [208].