Circulating endothelial (progenitor) cells reflect the state of the endothelium: vascular injury, repair and neovascularization
Abstract
An increase in the number of circulating endothelial cells (CEC) and of bone marrow-derived endothelial progenitor cells (EPC) in the peripheral blood is associated with vascular injury, repair and neovascularization. The phenotype and number of CEC may serve as diagnostic or prognostic parameters of vascular injury and tumour growth. An increase in the number of EPC may reflect repair of ischaemic vascular injury, a finding which has resulted in the initiation of clinical cardiovascular pilot trials using cell therapy. However, there is no consensus on the exact phenotype of the EPC and haematopoietic stem cells (HSC) and therefore the best candidate cell for transplant has not been established. Although the use of peripheral blood stem cells following mobilization, or of ex vivo-expanded cells, may improve EPC-mediated vascular graft endothelialization or tissue vascularization, sustained EPC-induced neovascularization still needs to be proven. Flow cytometric characterization, in combination with functional assays, will further elucidate the phenotype of the CEC and EPC, thereby providing reliable detection to appreciate their role in vascular diseases and cancer and to evaluate and, if possible, improve their therapeutic potential.
Physiological turnover of endothelial cells
The vascular network of fully grown adults is a dynamic organ with an estimated surface area of > 1000 m2. Located at the interface between blood and tissue parenchyma, the endothelial cell layer normally maintains an anticoagulant, antithrombotic and anti-inflammatory state. The structural and functional integrity of this network is maintained by continuous renewal of the endothelial cell layer, with a low basal replication rate of 0·1% per day. The endothelial cell responds to pathological conditions with rapid changes to an overt procoagulant and proinflammatory state. Normally, this activated state is functional, local and reversible, but by inappropriate or sustained activation or triggering of the endothelium, vascular diseases may develop [1]. In general, endothelial activation, e.g. by trauma, microbial organisms or toxins, also induces apoptosis and dislodgement of endothelial cells. This will result in elevated levels of circulating endothelial cells (CEC) and an additional demand for endothelial cell formation.
Vascular injury: endothelial cell detachment
The physiological turnover of the endothelium is reflected by low basal levels of CEC. In steady-state conditions, 99% of endothelial cells are quiescent. In healthy volunteers, hardly any CEC are detectable in the peripheral blood. Junctions with neighbouring cells, anchorage to the extracellular matrix (mediated by integrins, cadherins, fibronectin and vitronectin) and local shear stress are crucial for the homeostasis and survival of endothelial cells. In contrast, vascular stimulation or damage results in increased expression of tissue factor (TF), which can initiate thrombus formation. Up-regulation of several adhesion molecules on the endothelial cells, i.e. E-selectin and P-selectin (CD62E and CD62P), intercellular adhesion molecule-1 (ICAM-1/CD54) and vascular cell adhesion molecule-1 (VCAM-1), enables tethering, rolling and adhesion of leucocytes to the endothelium. Adhesion precedes the subsequent transendothelial migration of leucocytes to sites of inflammation. These pro-coagulant and pro-inflammatory processes are associated with increased detachment of endothelial cells. Several mechanisms of endothelial cell detachment are summarized in Table 1[2–8]. Two mechanisms are of particular interest. In the first, cytokines, especially tumour necrosis factor (TNF) and interferon (IFN), will cause proteolysis of the endothelial cell matrix (ECM) and thus disturb endothelial adherence. Cytokine (TNF and IFN) therapy is now used to inhibit tumour angiogenesis [8]. In the second, induction of endothelial apoptosis as the result of alterations in the balance of pro-apoptotic and antiapoptotic factors also results in endothelial cell detachment, although not all CEC are apoptotic, and this percentage differs among vascular diseases. Patients with sickle cell disease, and patients with acute coronary syndromes, showed a very low degree of apoptotic CEC. High plasma levels of vascular endothelial growth factor (VEGF) may be responsible for this, while in vitro, Solovey and colleagues showed that VEGF inhibits the apoptosis of cultured endothelial cells that were kept ‘unanchored and not allowed to re-establish attachment’[9]. On the other hand, in rickettsial infection, high levels of necrotic CEC have been reported [10]. As endothelial cells from diverse tissues and vascular beds are also heterogeneous with respect to (electron microscopic) appearance, functional capacity and surface phenotype and protein expression, characterization of CEC may give clues on both the pathogenic cause for cell detachment and of their vascular origin. For example, CEC found in sickle cell patients originate from microvessels, and in acute coronary syndromes from macrovessels [11,12]. Thus, by measuring the levels of CEC and by analysing their phenotypes, knowledge on the severity and pathogenesis of vascular diseases can be obtained in a direct and non-invasive manner.
Vasculogenesis in adult life
Until recently, it was thought that in adult life vessel formation was only mediated by angiogenesis and arteriogenesis, defined as the formation of new blood vessels by sprouting of endothelial cells from pre-existing blood vessels, and remodelling of arterioles (bypassing an obstruction) into larger size collaterals, respectively. Angiogenesis is important for general development, reproduction and wound healing, but also for tumour growth and metastasis. In contrast, the process of stem cell-mediated formation of blood vessels during embryonic development is called vasculogenesis. Emerging evidence suggests that stem cells might also contribute to vessel formation in adult life. Indications of postembryonic vasculogenesis have been demonstrated mostly in animal models, but pioneering human studies have recently also been conducted [13–16].
Animal studies
The first evidence for postnatal neovascularization by human peripheral blood-derived CD34+ precursors was shown in an immunodeficient mice model of hindlimb ischaemia by Asahara and colleagues, who described the incorporation of progenitors into sites of active angiogenesis [13]. They also suggested that these endothelial progenitor cells (EPC) might be useful for augmenting collateral vessel growth in ischemic tissues and for delivering anti- or pro-angiogenic agents. In another study, Shi and colleagues investigated the neovascularization of a synthetic graft of the thoracic aorta in a canine bone marrow transplantation model [14]. They found endothelial cells from donor origin, and therefore derived from the bone marrow graft, lining the prosthesis 12 weeks postangioplasty. Further convincing evidence came from Asahara and colleagues, who used transgenic mice in which endothelial cell-specific promoters, Flk-1 or Tie-2, regulate lacZ expression [15]. Following transplantation of transgenic marrow to normal animals, endothelial cells expressing lacZ were found in recipient vessels with a consistently fast turnover, i.e. in bone marrow, skin, postischaemic hindlimb muscle, ovary and uterus, but also in growing tumours.
Human studies
Rafii & Lyden have reviewed, in detail, preclinical and clinical pilot studies which confirm the animal data by showing the existence of in vivo postnatal vasculogenesis in humans [16]. These studies suggest a possible role for bone marrow, and thus stem cell-derived EPC, in physiological maintenance and repair. However, these studies do not prove that the clinical effects depend on the (generally limited) number of EPC that are part of the expanding and differentiating vascular cells in new blood vessels. The continuous functional presence of bone marrow-derived cells in newly formed blood vessels remains to be proven.
Origin and mobilization of haematopoietic stem cells (HSC) and of EPC
The presence of stem cells capable of forming various tissues in adult life is the holy grail of current medicine involved in tissue engineering. More than 30 years ago, adult stem cells were already the subject of investigation and extensively studied in chick-quail chimaeras for both haematopoietic and non-haematopoietic potency [17]. Accumulating evidence suggests the existence of a common precursor cell for both blood and endothelial cells in adult life. This precursor might be the adult equivalent of the haemangioblast that has been identified in embryonic development [18–20]. Additionally, a population of primitive cells was described that had an even larger multipotent differentiation potential. These cells were able to differentiate into mesenchymal cells, and into cells with neuro-ectoderm, endoderm and visceral mesoderm (i.e. endothelial cells) characteristics in vitro. These multipotent adult progenitor cells (MAPC), identified in long-term culture of human adherent bone marrow cells, confirm the existence of primitive cells in adult life [21,22]. Unfortunately, the isolation, characterization and culture of MAPC appear very troublesome and their availability for research is still very limited. Finally, tissue-specific stem cells have been reported, but in adulthood most stem cells reside in the bone marrow. For transplantation purposes, HSC are now primarily acquired from mobilized peripheral blood instead of bone marrow. In animal studies, stimulation by growth factors [granulocyte–colony-stimulating factor (G-CSF), granulocyte – macrophage colony-stimulating factor (GM-CSF), interleukin (IL)-2, IL-3, IL-6, erythropoietin (EPO), stem cell factor (SCF), IL-8 and IL-1β] that are particularly known to mobilize HSC, also results in the presence of an increased number of EPC [23]. In human research, corresponding findings have been described for several cytokines [6,24]. This conformity between EPC and HSC response might reflect the common origin of these cells (haemangioblast), or a physiological mechanism for the mobilization of progenitor cells from bone marrow. An attractive hypothesis for the concomitant mobilization of HSC and EPC is a physiological need of synergistic interactions between these cells in the process of angiogenesis and vasculogenesis, as HSC mobilization also occurs as response to vascular injury. In this respect, it is thought that VEGF-A, placental growth factor (PlGF) and stromal-derived factor (SDF), released by blood platelets and monocytes, activate metalloproteinase-9 (MMP-9), which mediates a joint mobilization of HSC, EPC and other cell types. The interactions between these cells and EPC, which may contribute to the revascularization process, are depicted in detail in Fig. 1.
Circulating endothelial cells (CEC), endothelial progenitor cells (EPC) and haematopoietic stem cells (HSC) in vascular damage, repair and neovascularization. Mobilization and recruitment of endothelial, lymphatic and haematopoietic (progenitor) cells is initiated by vascular trauma. Secondary released pro-angiogenic factors [vascular endothelial growth factor (VEGF-A), placental growth factor (PlGF)] activate metalloproteinase 9 (MMP-9), which results in an increase in soluble kit ligand (sKitL), enhancing the proliferation of progenitor cells. Neovascularization results from proliferation of endothelial cells at the site of injury (angiogenesis) and mobilization of HSC, EPC and progenitors from the bone marrow into the circulation and incorporation of EPC (vasculogenesis) at the neoangiogenic site. CEC may be detected in the peripheral blood after detachment caused by vascular injury. ECM, extracellular matrix.
Characterization of human CEC, EPC and HSC
The existence and culture of mature endothelial cells from peripheral blood has been possible since the 1970s. However, only in the 1990s did it become possible to reproducibly and selectively identify and isolate endothelial cells in whole blood by fluorescence- and magnetic-activated cell sorting (FACS, MACS). Agreement on the phenotypic differentiation of EPC, HSC and CEC, however, is still lacking as several markers are shared by these cells. EPC also share functional characteristics with HSC, namely the capacity for self-renewal and the ability to give rise to one or more types of differentiated progeny clones. For ‘candidate HSC’, the ultimate test for this capacity is selection by flow cytometry followed by transplantation in irradiated mice, i.e. immunocompetent and immunodeficient mice for murine and human stem cells, respectively. Subsequent repopulation proves the presence and functionality of ‘true HSC’. Other assays used for HSC characterization are in vitro assays – clonogenic assay and long-term culture on a stromal layer. A priori selection of HSC on the basis of expression of cell-surface molecules (e.g. CD34+) is useful for the isolation of stem cell-rich populations, but the precise characterization of ‘true HSC’ is still a matter of debate. Similar issues involve the EPC. To elucidate the ‘true’ phenotype of the EPC, in vivo assays have now become available, i.e. testing their potential to restore the vascularization of ischaemic tissues. These in vivo assays, however, also need a priori-sorted cells (i.e. ‘candidate EPC’) for input.
Phenotypic identification
Although we acknowledge the difficulties in the definition of CEC, EPC and HSC, we will try to summarize their phenotypic characteristics, thereby focusing on human cells (Table 2). EPC and CEC are mostly defined by the expression of CD34 and VEGFR-2 (KDR/Flk-1), although the CD34+ KDR+ fraction also comprises HSC, as shown by Ziegler and colleagues [25]. HSC and EPC both express CD133 and c-kit (CD117), in contrast to CEC. Furthermore, HSC express CD38, in contrast to EPC, and generally do not express VE-Cadherin or fibroblast growth factor receptor (FGFR). Burger and colleagues proposed the following phenotype for the EPC: CD34+ FGFR+ CD38+ VE-Cadherin+ c-kit+ CD31+ KDR+ CD133+, which is generally compatible with the definition of several other authors [13,25–27]. If necessary, distinction between EPC and CEC can be made by expression of CD146 on CEC and by CD133 on EPC [28]. Activated CEC may be distinguished by expression of CD105 (endoglin), the receptor for transforming growth factor-β1, which is a recognized regulator of angiogenesis.
| Cell type | Origin | Phenotypic markers |
|---|---|---|
| CEC | Mature endothelium | CD34 KDR CD146 VE-Cadherin TM vWF |
| EPC | Bone marrow; umbilical cord blood; (mobilized) peripheral blood; vascular parenchyma and organ-specific EPC | CD34 KDR CD133 CD117 |
| VE-Cadherin CD38 FGFR | ||
| HSC | Bone marrow | CD34 KDR CD133 CD117 |
- FGFR, fibroblast growth factor receptor; KDR, kinase-inserted domain containing receptor (VEGFR-2); TM, thrombomodulin; vWF, von Willebrand factor.
Peripheral blood cell measurements
The methodology for detecting CEC evolved from smears of peripheral blood to immunomagnetic isolation and subsequent counting by fluorescent microscopy. By combining CD146 isolation with a secondary staining and addition of a standard number of fluorescent beads, it now appears possible to detect levels as low as 20 CEC/ml of blood by flow cytometry (L. Vigneron, personal communication; BioCytex, Marseille, France). The number of CEC in the peripheral blood, reported in healthy individuals, generally varies between 0 and 10 cells/ml, i.e. 0–0·0001% of the total white blood cell count (WBC) (Table 3). The maximum number of CD34+ cells reported in healthy volunteers is 5000 cells/ml (i.e. ≤ 0·05% of the total WBC) [29]. CD34+ cells that also express CD133 and VEGFR-2, or FGFR-1, or VEGFR-3, a population possibly enriched with EPC, correspond to 0·2–5% of all CD34+ cells (i.e. ≤ 0·0025% of the total WBC) [26,27,30]. In view of the low number of EPC and CEC within the total number of CD34+ cells, routine measurement of HSC is performed by flow cytometry on CD34 expression only [31,32].
| Disorder | Date | Mean CEC/ml | Cell markers | Ref. | |
|---|---|---|---|---|---|
| Patient | Control | ||||
| Coronary angioplasty | 1992 | 11 | 1 | CD146 TM vWF | [36] |
| Rickettsial infection | 1993 | 162 | NR | CD146 TM vWF | [10] |
| Sickle cell anaemia | 1997 | 2·6 | CD146 KDR Flt-1 TM | [11] | |
| Steady state | 13 | ||||
| Acute painful crisis | 23 | ||||
| Unstable angina | 1999 | 4·5 | 0 | CD146 CD36 vWF | [12] |
| Acute myocardial infarction | 1999 | 16 | 0 | CD146 CD36 vWF | [12] |
| 2004 | 4·9 | 1·0 | CD146 CD31 vWF | [34] | |
| ANCA-associated vasculitis | 2003 | 136 | 5 | CD146 CD31 UEA-1 | [38] |
| Stable coronary disease | 2004 | 18 | 10 | CD146 KDR CD31 CD106 Annexin-V | [35] |
| Peripheral atherosclerosis | 2004 | 1·0 | CD146 CD31 vWF | [34] | |
| Ischaemic rest pain | 3·5 | ||||
| Stable claudication | 1·1 | ||||
| SLE | 2004 | 89 | 10 | CD146 KDR CD31 CD106 Annexin-V | [35] |
| Allogenic stem cell transplantation | 2004 | 44 | 8 | CD146 UEA-1 | [39] |
| Cancer | 2004 | 399 | 121 | CD146 KDR CD31 vWF | [43] |
| Stable disease | 179 | ||||
| Progressive disease | 438 | ||||
- Flt-1, Fms-like tyrosine kinase 1 (VEGFR-1); KDR, kinase-inserted domain containing receptor (VEGFR-2); SLE, systemic lupus erythematosus; TM, thrombomodulin; UEA-1, Ulex europaeus agglutinin I; vWF, von Willebrand factor.
Culture and functional profiling
In parallel to HSC identification, definite proof of the presence of EPC, instead of merely CEC, requires culture of cells after isolation. In these functional assays, differences in growth kinetics become obvious for early EPC and differentiated CEC. Cells that enable late endothelial outgrowth (CFU-EC) might represent the ‘true’ angioblast-like EPC with high proliferative potential, while endothelial cell colonies that appear early in culture might more preferentially originate from CEC [33]. In conclusion, identification, isolation and purification of EPC and CEC are troublesome, which hampers the interpretation of and comparison with previous studies. More specific endothelial markers and validated functional assays are needed to confirm the characteristics of the isolated endothelial cell.
Circulating endothelial cells in vascular disorders
Elevated levels of CEC are found in a variety of conditions characterized by vascular injury or vessel formation (Table 3). In all studies, the number of CEC correlated with the degree of endothelial injury or neovascularization. Furthermore, CEC levels were higher during the acute phase of the clinical syndrome and were found to predict severity and outcome in vascular disease. For example, sickle cell patients in steady state had significantly fewer CEC than those in acute painful crisis, although their CEC levels were still elevated when compared to those of normal subjects [11]. In acute coronary syndromes, in sickle cell patients and in patients with systemic lupus erythematosus (SLE), elevated plasma TF levels correlated with high levels of CEC and may represent excessive endothelial activation associated with a prothrombotic state [34–36]. In young women with SLE, an increased number of apoptotic CEC was found that correlated with endothelial dysfunction or damage [35]. In Behçet's disease, in anti-neutrophil cytoplasmic autoantibody-associated vasculitis and following allogenic stem cell transplantation, elevated CEC levels were also reported [37–39]. In infectious diseases, CEC may serve as a carrier for micro-organisms and viruses. In cytomegalovirus (CMV) infections in immunocompromised hosts, isolated CEC were found to contain virus, and levels of CMV viraemia and antigenaemia in patients with acquired immune-deficiency syndrome (AIDS) or kidney transplants correlated with CEC numbers [40,41]. Treatment of CMV infection with ganciclovir or foscarnet resulted in the disappearance of CEC in the peripheral blood. CEC levels also reflect the effect of therapeutic interventions in other diseases with vascular damage. In thrombocytopathic thrombocytopenic purpura (TTP), a concomitant decrease in CEC and increase in number of platelets was shown after the institution of plasmapheresis and vincristin therapy [42]. Breast cancer and lymphoma patients achieving complete remission showed a reduction in CEC as compared to healthy individuals [5]. Other studies have confirmed the relationship between increased levels of CEC in cancer patients and progressive disease [43,44]. Measurement of tumour angiogenesis was previously based on the evaluation of mean vessel density (MVD) of the tumour. As correlation of MVD with clinical outcome is still uncertain in most tumour types, measurement of CEC numbers might represent an alternative, non-invasive, method to determine tumour angiogenesis [45].
Human research: EPC and HSC to the rescue
Until now, no systematic studies exist regarding physiological variation in the level of EPC, which is mainly caused by the absence of a consensus on the phenotype, as well as by difficulties in measuring such low numbers of cells in the circulation. In the currently available human studies, increased levels of EPC were reported in patients who experienced coronary artery bypass grafting (CABG), and burn injuries or myocardial infarction, as measured by KDR+ CD133+ VE-Cadherin+ cells and CD34+ cells, respectively [6,46]. Although this appears to confirm the joint mobilization of EPC with HSC by ischaemia-dependent release of pro-angiogenic factors, in other patients, i.e. in those with gastric and breast cancer similarly prone to this process, no elevated levels of circulating EPC were found [47]. Unfortunately, many reports on EPC levels do not incorporate EPC cultures following flow cytometric assessment. Hill and colleagues, however, showed an increased number of colony-forming units of endothelial progenitor cells in patients with increased cardiovascular risk, suggesting EPC-mediated repair under these conditions [48]. In another study, the number of EPC, as well as their migratory activity, was decreased in patients at risk for coronary artery disease, suggesting that impaired EPC-mediated vascular repair may even partly cause vascular disease [49]. Although the interpretation of in vivo results remains difficult, in several human studies, the incorporation of EPC in mature endothelium has been detected. In patients with end-stage heart disease, insertion of a left ventricular device results in the recruitment of progenitor cells to form a vascular surface [50]. EPC contribute to endothelial repair in thrombotic microangiopathy and vascular repair in kidney graft rejection [51,52]. As a consequence of the findings in animal and human studies, as well as in vitro assays, cell therapy for vascular diseases was initiated. The first results of clinical trials were published in 2002 (Table 4). Intramuscular administration of bone marrow mononuclear cells was found to be beneficial in peripheral artery disease in a randomized controlled trial [53]. Ischaemic myocardium showed improvement in stroke volume and in contractility and cardiac function by intramyocardial injection or intracoronary infusion of bone marrow-derived cells [54–57]. The safety and feasibility of intracoronary infusion of stem/progenitor cells after acute myocardial infarction was also demonstrated [58–60]. In a small randomized trial, the global left-ventricular ejection fraction (LVEF) did not change in the control group (no intervention), but significantly increased in the intervention group (intracoronary bone marrow infusion), i.e. from 50 to 56·7% after 6 months of follow-up [60]. Many studies, however, have failed to include a control population, and compare patients with historical controls. Furthermore, owing to the presence of HSC, EPC, MSC (multipotent stem cells) and other ‘facilitating cells’ in the bone marrow, it is unclear which cell or mechanism is responsible for the observed (beneficial) effect. In conclusion, these preliminary, preclinical studies show promising results, but much more research is necessary to appreciate the improvements in clinical outcome. This may be accomplished by including proper controls, by tracking the injected cells and following their fate in vivo, and by studying the long-term effects.
| Disorder | Cell type | No. of patients | Application | Controls (no. of controls) | Effects | Ref. |
|---|---|---|---|---|---|---|
| PAD | BMC | 45 | Intramuscular | Auto-control contralateral leg, saline (23)/PBMC (22) | ABI↑, transcutaneous oxygen pressure↑ rest pain↓, pain-free walking time↑ | [53] |
| HF | BMC | 10 | Intramyocardial | NA | Angina score↓, stress-induced ischemia↓ | [54] |
| BMC | 8 | Intramyocardial | NA | Angina episodes↓, ventricular wall motion and thickening↑ | [55] | |
| BMC | 14 | Intramyocardial | No infusion (7) | Ejection fraction↑, end-systolic volume↓ | [56] | |
| BMC | 6 | Local injection at CABG | NA | Ejection fraction↑ (n = 4) perfusion↑ (n = 5) | [57] | |
| AMI | BMC | 10 | Intracoronary | No infusion (10) | Hypokinetic area↓, end-systolic volume↓ contractility↑, perfusion↑ | [58] |
| BMC | 9 | Intracoronary | No infusion (11) | Hypokinetic area↓, end-systolic volume↓, contractility↑, perfusion↑ | [59] | |
| EPC | 11 | |||||
| BMC | 30 | Intracoronary | No infusion (30) | Ejection fraction↑ | [60] |
- ABI, ankle-brachial index; AMI, acute myocardial infarction; BMC, bone marrow-derived cells; CABG, coronary artery bypass graft; EPC, endothelial progenitor cells; HF, chronic ischaemic heart failure; NA, not applicable; PBMC, peripheral blood mononuclear cells; PAD, peripheral artery disease.
Conclusions, considerations and future directions
CEC and EPC may serve as diagnostic, therapeutic or prognostic parameters of vascular injury and neovascularization and provide new insights into the pathobiology of the involved endothelium and the successive bone marrow response. However, many issues still need to be resolved in endothelial (progenitor) cell research and therapy. Lack of consensus on definitions of CEC and EPC hampers longitudinal in vivo identification of CEC, EPC and other bone marrow-derived cells. Therefore, the appreciation of their contribution to formation of blood vessels in humans remains troublesome. Only if consensus is reached on the most effective target cells, or combination of cells, does comparison between studies become possible and permit dose–effect studies with proper controls. The first preclinical studies addressing EPC-mediated vasculogenic therapy have recently been conducted. So far, these studies have shown only a minor contribution of EPC to vessel formation, and optimizing of EPC-mediated therapy is necessary. The joint mobilization of EPC and HSC in peripheral blood offers possibilities for collection and subsequent therapeutic use of EPC, but isolation, culturing and ex vivo expansion may be unavoidable to obtain sufficient cell numbers. Alternatively, amplification of the synergy between EPC and growth factor-producing blood cells (e.g. monocytes and platelets), and augmentation of classical (local) angiogenesis should also be considered. Stimulation of adhesion and incorporation of cells at sites where neovascularization actually takes place, i.e. targeting of the homing process, which will strongly depend on the route of administration, should also be the subject of systematic studies. Finally, inhibitory therapies should be considered if EPC-mediated vasculogenesis plays a role in tumour growth and vascular proliferative diseases, but this also remains to be proven.
