Published ahead of print on June 23, 2005, doi:10.1164/rccm.200410-1325OC
© 2005 American Thoracic Society doi: 10.1164/rccm.200410-1325OC
Increased Circulating Endothelial Progenitor Cells Are Associated with Survival in Acute Lung InjuryDivisions of Pulmonary, Allergy, and Critical Care, and Cardiology, Department of Medicine, Emory University School of Medicine; and the Atlanta Veterans' Affairs Medical Center, Atlanta, Georgia Correspondence and requests for reprints should be addressed to Ellen L. Burnham, M.D., Grady Memorial Hospital, 69 Jesse Hill Jr Drive SE, Suite 2D, Atlanta, GA 30335. E-mail: eburnha{at}emory.edu
Rationale: Repair of damaged endothelium is important in recovery from acute lung injury. In animal models, bone marrow–derived endothelial progenitor cells differentiate into mature endothelium and assist in repairing damaged vasculature. Objectives: The quantity of endothelial progenitor cells in patients with acute lung injury is unknown. We hypothesize that increased numbers of circulating endothelial progenitor cells will be associated with an improved outcome in acute lung injury and the acute respiratory distress syndrome. Methods: Peripheral blood mononuclear cells from the buffy coat of patients with early acute lung injury (n = 45), intubated control subjects (n = 10), and healthy volunteers (n = 7) were isolated using Ficoll density gradient centrifugation, and plated on fibronectin-coated cellware. After 24 hours, nonadherent cells were removed and replated on fibronectin-coated cellware at a concentration of 1 x 106 cells/well. Colony-forming units were counted after 7 days' incubation.
Measurements/Main Results: Endothelial progenitor cell colony numbers were significantly higher in patients with acute lung injury compared with healthy control subjects (p < 0.05), but did not differ between patients with acute lung injury and intubated control subjects. However, in the 45 patients with acute lung injury, improved survival correlated with a higher colony count (p < 0.04). Patients with acute lung injury with a colony count of Conclusions: An increased number of circulating endothelial progenitor cells in acute lung injury is associated with improved survival.
Key Words: acute respiratory distress syndrome • endothelium • mortality • outcomes • stem cells The acute respiratory distress syndrome (ARDS) and acute lung injury (ALI) are devastating causes of respiratory failure characterized by disruption of the alveolar–capillary membrane, resulting in accumulation of proteinaceous pulmonary edema fluid and coincident hypoxemia. Patients with ALI/ARDS usually require mechanical ventilatory support, have a prolonged length of stay in an intensive care setting, and continue to have an unacceptably high mortality, ranging between 30 and 50%. Despite recent advances in therapeutic ventilatory strategies for patients with ARDS, efforts to identify circulating factors that predict survival in these critically ill patients have been unrevealing (1). In addition, a full explanation of the mechanisms responsible for repairing the injured alveolar–capillary membrane in patients with this disorder remains elusive. Stem cells have been identified in adult humans that are capable of maintaining, generating, and replacing terminally differentiated cells during normal physiologic turnover, and in response to acute tissue damage (2, 3). Bone marrow–derived stem cells can differentiate into a variety of tissue-specific, human cell types (4–7). In a murine model, alveolar and bronchial epithelial cells derived from bone marrow donors have been identified after stem cell transplant in lethally irradiated recipients (8). In humans, after allogeneic hematopoietic stem cell transplantation, pulmonary endothelial and epithelial cells have been observed in the lungs of recipients that are of donor origin (9). Endothelial progenitor cells (EPCs) are a specific subtype of hematopoietic stem cell that has been isolated from the peripheral blood of humans (10–12). EPCs migrate from the bone marrow to the peripheral circulation where they contribute to the repair of injured endothelium and to the formation of new blood vessels (10, 13). Hill and colleagues (14) have reported that levels of circulating EPCs may be a prognostic biological marker for vascular function and cumulative cardiovascular risk in male subjects without a history of cardiovascular disease. However, circulating EPCs have not previously been identified in critically ill patients. Damage to the pulmonary vascular endothelium occurs in patients with ARDS (15). EPCs may provide a circulating pool of cells that could form a cellular "patch" at the site of denuding injury, or could serve as a reservoir to replace damaged endothelium (14). It is unknown at present if circulating progenitor cells are associated with enhanced repair of damaged lung, including the pulmonary endothelium, in patients with ALI. We hypothesized that EPCs are present at increased levels in the circulation of patients with ALI, and that measurement of the number of circulating EPCs may serve as a prognostic biomarker for patients with ALI/ARDS. Some of the results of this study have been previously reported in the form of an abstract (16).
Additional detail on the methods is provided in an online supplement.
Patient Characteristics
EPC Colony Isolation Patients with ALI who remained in the ICU on a ventilator had a second blood specimen obtained on Day 7, which was processed identically.
Analysis
Demographics of Patients with ALI and ICU Control Patients Forty-five patients with ALI and 10 ICU control patients were enrolled in the protocol (Table 1). The patients were similar demographically except for an older age of the ICU control patients (53 years in patients with ALI [25–75% quartiles, 42–64 years] vs. 63 years in ICU control patients [25–75% quartiles, 60–69 years], p < 0.03), and significantly higher Acute Physiology and Chronic Health Evaluation (APACHE) II (p < 0.001) and Sequential Organ Failure Assessment (SOFA) scores (p < 0.001) in the patients with ALI. In addition, the degree of hypoxemia in the patients with ALI was greater than in ICU control patients (p < 0.001), and there was a slightly lower percentage of monocytes (p < 0.03). The number of cells obtained from the supernatant after the initial preplating step that were actually plated for the final CFU assessment was similar between the two groups (p = 0.16). Seven healthy volunteers without any comorbid diagnoses were also enrolled, with a mean age of 32 ± 2 years.
The risk factor for ALI was secondary to direct lung injury, predominantly pneumonia, in 33 of the 45 patients with ALI. Septic shock was present in 44% of the patients with ALI. The overall mortality rate for the patients with ALI was 42%. Comparing the survivors with the nonsurvivors of ALI, both groups had direct lung injury as a predisposing factor for the development of ALI, with pulmonary infections predominating (Table 2). The incidence of sepsis and septic shock defined using established criteria (19) was also not statistically different between survivors and nonsurvivors. The survivor and nonsurvivor subgroups were similar in terms of age, sex, and alcohol and tobacco use, and the number of cells available after the initial preplating step that were actually plated for the final CFU assessment was similar between the two groups (p = 0.61). Survivors had lower initial APACHE II and SOFA scores, and a small but significantly higher peripheral white blood cell count (p < 0.01).
The overall mortality rate for the ICU control patients was 20%. These patients had a variety of underlying diagnoses and the majority were ventilated for airway protection (Table 3).
Morphology of Progenitor Cell Colonies A photomicrograph of a typical EPC CFU on the fibronectin-coated plate from a patient with ALI is illustrated in Figure 1a, and a representative ICU control patient's CFU is illustrated in Figure 1b. Morphologically, the CFUs from our patients and control subjects (both ICU and healthy) appeared similar to those obtained by us and other investigators using these isolation methods (14, 20). In a subset of 13 consecutive patients' samples, we assessed the ability of the cells in the CFUs to incorporate acetylated low density lipoprotein (LDL) and to bind endothelial-specific lectin via immunostaining. We demonstrated that cells in these patients' colonies stained for both of these markers, consistent with what has been reported in the literature to represent an endothelial lineage (Figure 2) (10, 12, 14).
EPC Colony Numbers in Entire Cohort of Patients with ALI and ICU Control Patients The absolute numbers of CFUs per well present in the patients with ALI, in the ICU control patients, and in ambulatory control subjects are illustrated in Figure 3. The patients with ALI had a significantly higher number of CFUs than the healthy control subjects (p < 0.05), but the median number of CFUs was not statistically different between the patients with ALI and ICU control patients (p = 0.95). The number of CFUs in the healthy control subjects was similar to previously reported values for healthy individuals in the literature using this method of isolating EPCs (20).
Relationship of EPC Colony Numbers to Demographic Factors and Survival In the subset of patients with ALI, there was a significant survival advantage in the ALI patients with a higher number of EPC CFUs per well (p < 0.04; Figure 4). The number of CFUs per well did not correlate with a variety of demographic factors and variables that have previously been associated with outcome, including the following: the cause of ALI (direct vs. indirect lung injury; p = 0.97), degree of hypoxemia (r2 = 0.02, p = 0.32), age (r2 =0.01, p = 0.5), tobacco use (p = 0.24), alcohol abuse (p = 0.45), peripheral white blood cell count (r2 = 0.07, p = 0.08), number of cells plated (r2 = 0.04, p = 0.10), APACHE II score (r2 = 0.05, p = 0.14), or SOFA score (r2 = 0.08, p = 0.07). The number of CFUs per well did correlate positively with male sex (p < 0.03). The EPC CFU number did not significantly differ based on whether or not a patient was septic or had septic shock. The median EPC CFU number for nonseptic patients (n = 4) was 71 (25–75% confidence interval, 28–198), compared with septic patients (n = 21) who had a median of 114 (range, 23–205) CFUs, and patients with septic shock (n = 20) who had a median of 61 (range, 18–193; p = not significant for all comparisons).
After examining the receiver operator characteristics of our data, an EPC CFU number of 35 was used to stratify patients with ALI to maximize the sensitivity and specificity of EPC CFUs in the prediction of survival. Of the patients, 60% (27 of 45) had EPC CFUs exceeding 35, and 40% (18 of 45) had CFUs less than 35. When the patients with ALI were stratified based on this number, hospital mortality was significantly higher in the group with CFUs below 35 (61 vs. 30%, p < 0.03). A Kaplan-Meier curve illustrating the percentage of patients surviving stratified by an EPC CFU count of greater than or less than 35 is illustrated in Figure 5. At 28 days, approximately 75% of the patients with ALI with EPC CFUs greater than or equal to 35 were alive, compared with only 35% of the patients with EPC CFUs less than 35 (p < 0.03). In a multivariable analysis controlling for the effects of age, sex, and severity of illness (SOFA scores), the association between EPC CFUs greater than 35 and survival remained, with an odds ratio of 4 (p = 0.06; 95% confidence interval, 1–20). No significant difference in the median number of EPC CFUs was observed in those ICU control patients who survived and those who expired, although the number of these patients was small (n = 10).
EPC Colony Numbers over Time in Patients with ALI From our initial cohort of 26 patients, 17 patients with ALI consented to a second blood draw 7 days after the initial specimen. Blood was not obtained in 9 of these 26 patients with ALI because 6 had expired, 2 were no longer ventilated, and 1 refused a subsequent blood draw. In this subgroup of patients, there were no significant changes in the EPC CFUs over time (Figure 6). At 7 days after the initial sampling, the ALI survivors (n = 12) had a higher median number of EPC CFUs (115, with 25–75% quartiles of 34–150) compared with ALI nonsurvivors (n = 5) (55, with 25–75% quartiles of 22–151), although this was not statistically significant. The additional 19 enrolled patients with ALI did not provide consent for subsequent blood sampling given the stability of the EPC CFU numbers over time in the initial ALI cohort.
This study demonstrated that EPCs are mobilized in the circulation of patients with ALI and ARDS, disorders characterized by severe pulmonary endothelial injury. Overall, the number of circulating EPCs in patients with ALI is approximately twofold higher than in healthy control subjects. More importantly, an increased number of circulating EPCs was associated with improved survival, even after correcting for differences in age, sex, and severity of illness. In those patients with ALI with blood specimens obtained over a 7-day period, the number of circulating EPCs did not significantly change. Thus, the measurement of EPC CFUs is promising as a novel prognostic marker for survival in patients with ALI. These findings also support the possibility that adequate mobilization of EPCs from the bone marrow in ALI could potentially contribute to repair and recovery of damaged pulmonary endothelium. Recent studies in mice lend credence to the idea that marrow-derived stem cells can serve as progenitor cells for the tissues of various solid organs (21). Studies in the lung have focused mainly on the regeneration of pulmonary epithelium. After transplanting a single marrow stem cell into recipient mice that had been marrow-ablated, Krause and colleagues (8) demonstrated engrafted donor cells with alveolar type II cell and airway epithelial cell phenotypes. Engraftment in the lungs of these animals was the highest of any organ studied. In a study by Kotton and colleagues (22), plastic-adherent lacZ+ marrow cells were injected into mice without marrow ablation, and apparent engraftment of cells bearing type I pneumocyte morphology was observed, which was particularly marked after pretreatment with intratracheal bleomycin. Furthermore, mesenchymal stem cells are able to engraft in the lung after bleomycin injury. These cells originated from the bone marrow, and after engraftment exhibited an epithelial morphology (23). Recently, the derivation of type I alveolar cells, lung fibroblasts, and interstitial monocytes/macrophages from circulating progenitor cells has been reported in a parabiotic mouse model, most robustly in conjunction with irradiation and elastase exposure (24). A common conclusion of all these studies is that the presence of ALI in the transplant recipients (either from radiation or via chemical agents) facilitates engraftment and presumably the stem cell–dependent repair process. Less has been reported regarding repair of damaged pulmonary endothelium, although pulmonary endothelial damage is characteristic of a variety of conditions, including ALI (15). The utility of circulating EPCs as a treatment for damaged vasculature has been reported in ischemic vascular disease (25) and for the treatment of ischemic retinopathy (26). A similar type of therapy for damaged pulmonary endothelium appears possible. Transplanted hematopoietic stem cells home to the pulmonary vasculature, as reported by Suratt and colleagues (9). These investigators examined lung biopsy specimens from a group of female patients who had undergone cord blood transplantation from male donors. A modest rate of epithelial chimerism (2.5–8.0%) was observed in the recipients' lungs in addition to a more marked rate of endothelial chimerism (37.5–42.3%), promoting the idea that adult human progenitor cells could play a potentially therapeutic role in lung repair. Moreover, the utility of EPCs may not be limited to these cells merely replacing the injured vascular endothelium. For example, EPCs have been reported to serve as a potential source for angiogenic factors (27). They may secrete paracrine growth factors that regulate the angiogenic response and may participate in cell–cell contact events that facilitate endothelial cell sprouting and growth. EPCs may be useful for disease risk stratification and prognosis, as illustrated by Hill and colleagues (14) where patients with lower numbers of circulating EPCs had a higher Framingham risk score, signifying a greater risk for cardiac events. These investigators were also able to demonstrate a relationship between subjects' endothelial function (as measured by brachial reactivity) and the number of circulating EPCs. Circulating EPCs in our patients with ALI presumably originate from the bone marrow, suggesting novel therapeutic interventions to increase the numbers of these progenitor cells. Mobilization of hematopoietic stem cells from the bone marrow to the periphery has been induced clinically or experimentally through the use of cytokines, chemokines, and chemotherapeutic agents, and this has potential therapeutic implications for enhancing lung repair in ALI. The most commonly used strategy in humans to mobilize stem cells is the use of granulocyte colony–stimulating factor or granulocyte-monocyte colony–stimulating factor, usually in conjunction with disease-specific chemotherapeutic agents (28). Importantly, granulocyte-monocyte colony–stimulating factor has been demonstrated to improve gas exchange in a small series of patients with severe sepsis and respiratory dysfunction (29). However, the role of granulocyte-monocyte colony–stimulating factor in the treatment of ALI remains to be determined. A majority of the patients with ALI in this study were septic. Recent investigations have examined lethally irradiated C57Bl/6 mice that have been reconstituted with bone marrow–derived from green fluorescent protein-transgenic mice. The mice were exposed to either intranasal LPS (found in the bacterial capsule of gram-negative organisms) or to intranasal saline. A significantly increased number of circulating EPCs was present in the circulation of the LPS-exposed animals compared with the saline-treated animals (30). An elevation in the circulating number of mature endothelial cells has been reported in patients with sepsis, a major risk factor for the development of ALI, indicative of heightened endothelial cell activation (31). These works highlight potential mechanisms involving inflammatory mediators in endothelial damage and repair. We did not measure the level of endothelial cells in our patient population and a relationship between EPCs and endothelial cells remains to be evaluated. The number of EPC CFUs in our patients with ALI correlated positively with male sex. In animal models, estrogens have been reported to have antiapoptotic effects on progenitor cells (32). Given that the median ages of our patients with ALI was in the perimenopausal range, the women in the cohort would be expected to have lower estrogen levels than a younger cohort of patients. Lower estrogen concentrations in this subgroup of women with ALI may have been operative in lowering the values of EPC CFUs in these patients compared with men. The number of men in the ALI survivor and nonsurvivor groups was similar, however. Our study is not without potential limitations. EPCs were identified in these patients' circulation using methods that have been previously described and validated by us and others (14, 20). EPC CFU numbers in our healthy control patients were similar to those reported previously (20). Nevertheless, some degree of contamination by non-EPCs is possible. To minimize this problem, the initial preplating of the peripheral blood mononuclear cells was performed to avoid isolating mature circulating endothelial cells. In addition, we used markers of endothelial lineage to further clarify the type of cells growing in culture. Although these methods are widely accepted, they are still not perfectly accurate because few markers and stains are specific for a cell type, and results may vary depending on the immaturity or senescence of the cell (33). Factors such as timing of the blood sampling in relationship to the onset of lung injury did not have an immediate impact on the colony number, as counts remained stable over a week in patients with ALI. Finally, this study did not use other functional assays previously performed with EPCs to assess migration and proliferation. However, we believe that the cell-culturing technique we used in our assay incorporates a measure of not only the number of circulating EPCs but also the ability of these EPCs to form endothelioid colonies, thus encompassing both the migratory and proliferative capacity of this progenitor cell population. The observation that our ICU control patients had similarly elevated EPCs was intriguing but not completely unexpected. These patients were critically ill on mechanical ventilation, although their severity of illness was less than in subjects with ALI. It is unknown at this time what an appropriate ICU control patient should be, although this group of control patients met no criteria for ALI, and had no risk factors for its development. The group of ICU control patients was small (n = 10), and had a preponderance of neurologic diagnoses, including cerebrovascular events, which may have resulted in elevations in circulating EPCs (34). Further investigation in a larger number of non-ALI control subjects is necessary to demonstrate the prognostic significance of EPC number in critically ill patients with other diagnoses. In summary, the measurement of circulating EPCs appears to have prognostic implications for ALI, a disorder characterized by severe endothelial damage. Pulmonary and critical care investigators have expressed interest in the role of stem cells to help repair the damaged lung in ALI, with the understanding that mechanisms of stem cell release, targeting, proliferation, and differentiation in lung tissue must be established (35). Defining the role of EPCs in repair of the acutely injured lung will provide a rationale for novel approaches to therapy, but the utility of stem cell therapy as a clinical modality requires additional basic and clinical research.
The authors thank Meredith Mealer, Leslie Rogin, Marsha Burks, and Ginny Wiggs for assistance with patient enrollment, and Diane Sutcliff and Patrick Cowan for their technical assistance.
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org Conflict of Interest Statement: None of the authors have a financial interest with a commercial entity that has an interest in the subject of this manuscript. Received in original form October 6, 2004; accepted in final form June 22, 2005
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
35 had a mortality of 30%, compared with 61% in those with colony counts < 35 (p < 0.03), results that persisted in a multivariable analysis correcting for age, sex, and severity of illness. 
value of 0.05 was used in all analyses. 
