 |
||||
| 2007 JASN IMPACT FACTOR 7.111 | HOME AUTHOR INFO EDITORIAL BOARD SUBSCRIBE FEEDBACK ALERTS HELP | |||
| CURRENT ISSUE | ARCHIVES | JASN Express | ONLINE SUBMISSION | |
|
||||
| 2007 JASN IMPACT FACTOR 7.111 | HOME AUTHOR INFO EDITORIAL BOARD SUBSCRIBE FEEDBACK ALERTS HELP | |||
| CURRENT ISSUE | ARCHIVES | JASN Express | ONLINE SUBMISSION | |
Nephrology beyond JASN |
Address correspondence to: Prof. Eberhard Ritz, Department Internal Medicine, Division of Nephrology, Bergheimer Strasse 56a, D-69115 Heidelberg, Germany. Phone: +49-0- 6221-601705 or +49-0-6221-189976; Fax: +49-0-6221-603302; E-mail: Prof.E.Ritz{at}t-online.de
Sepsis is a frequent medical emergency. Despite much recent progress in patient management, it still carries a devastating prognosis (1). Sepsis with multiorgan failure is well known to nephrologists because of its frequent association with acute renal failure (2,3). The pathomechanisms involved are becoming ever more complex and have not been completely elucidated (4,5). The extravasation of fluid into lung tissue causing the respiratory distress syndrome ("white lung") is one dire facet of multiorgan failure and continues to be a frequent and major cause of death (6,7). There is a desperate need for innovative interventions and the search goes on (8).
The work of Parikh et al. (9), if confirmed, promises to be a major breakthrough that may open new perspectives for focused and rational therapies.
In a well-organized series of studies from bedside to bench, the authors examined septic patients with one major complication of sepsis (i.e., impaired oxygenation as a result of pulmonary vascular leak), identified angiopoietin-2 as a potential culprit, then went to the bench conducting in vitro and in vivo studies to verify that angiopoietin-2 did indeed cause vascular leak, and finally attempted to identify the underlying molecular mechanism(s). The results of this logical sequence are analogous to fulfilling Koch’s postulates.
What is angiopoietin-2? For the nonexpert, some background information may be useful (10). The formation of vessels in the embryo begins when angioblasts differentiate into endothelial cells and assemble into tubes in response to the signals of a large family of vascular endothelial growth factor (VEGF) isoforms from surrounding tissues (11). Tissues "speak" to endothelial cells by secreting VEGF (12). The importance of VEGF in angiogenesis is illustrated by the fact that the loss of two alleles with VEGF is lethal and the loss of one allele causes vascular defects. VEGF interacts with two receptor kinases: VEGF-R1 or flt, and VEGF-R2 or flk (13). The former is well known to nephrologists, because a soluble isoform acts in the circulation as a decoy receptor scavenging and inhibiting the action of proangiogenic VEGF. It has recently been identified as one cause of preeclampsia (14).
Once VEGF has caused endothelial cell induction, the further downstream steps in angiogenesis are a highly complex, coordinated process through the sequential action of a series of receptors and numerous ligands (10,11).
In microvessels, PDGF is involved in recruiting pericytes. In larger vessels, it is primarily the angiopoietin-1/Tie2 (tyrosine kinase with Ig and EGF homology domains) ligand/receptor pair that is involved in recruiting smooth muscle cells.
An important recent concept is the notion that maintenance of vascular integrity is the result of a balance between stabilization and regression. Although endothelial cells have half-lives of several years, without the input of survival signals that stabilize endothelial cells and smooth muscle cells, vascular regression and disintegration will occur.
It is here that the angiopoietins enter the story: the Tie2 receptor binds angiopoietin-1 and angiopoietin-2. Angiopoietin-1 antagonizes the proangiogenic and permeability-promoting cytokine VEGF. Angiopoietin-1 is an endothelial survival factor. It causes vascular maturation and quiescence. It is ubiquitously expressed and preserves cell-to-cell contacts (15). It phosphorylates the Tie2 receptor and is usually antiangiogenic, since it stabilizes and tightens up the vessel wall.
In contrast, the antagonistic ligand angiopoietin-2 is not ubiquitously expressed. Notably, however, it is strongly expressed in the lung (16) together with the Tie2 receptor (17). Angiopoietin-2 initiates vascular remodeling by loosening up the vessel wall, a precondition for vascular remodeling—and, as one could predict, such loosening up might also increase vascular permeability. Furthermore, angiopoietin-2 has functions beyond vascular morphogenesis: it is stored in Weibel-Palade bodies, from where it is released rapidly upon stimulation, e.g., by proinflammatory signals, thus initiating rapid vascular responses (18).
The involvement of the angiopoietin-1/angiopoietin-2 system in septic shock had been suggested by the observation of Witzenbichler et al. (19) that angiopoietin-1 protected mice with endotoxin shock: it prevented hemodynamic instability, reduced lung water content, and diminished myeloperoxidase activity. Baffert et al. showed that this was mediated by a reduction in the number and the size of endothelial gaps (20). Conversely, it has been shown that angiopoietin-2 is crucial for the inflammatory vascular response to chemicals or bacterial infection sensitizing endothelial cells to TNF
(18).
In the study highlighted here, Parikh et al. (9) examined whether angiopoietin-2 played a role in the respiratory distress syndrome of septicemia. In a small sample of 22 patients with sepsis and 29 control patient in the general medical service, the authors prospectively sampled serum and measured angiopoietin-2 concentrations, which were significantly higher in patients with sepsis and multiorgan dysfunction (23.2 ± 9.1 ng/ml) than in nonseptic controls (3.5 ± 0.6 ng/ml) or septic patients without multiorgan dysfunction (4.8 ± 1.5 ng/ml).
Next, Parikh et al. postulated that the high expression of angiopoietin-2 (16) and Tie2 receptor (17) in the lung made it a plausible target organ for angiopoietin-2–mediated organ dysfunction. Therefore they examined the relation between angiopoietin-2 concentrations and indicators of poor oxygenation (PaO2/FIO2). Indeed, a strong correlation was found.
Advancing from bedside to bench, the authors next studied whether the serum of septic patients disrupted the architecture of human microvascular endothelial cells obtained from neonatal dermis. This tested the hypothesis that actin-myosin–driven cell contraction opens gaps between endothelial cells, thus permitting paracellular escape of molecules across the endothelial barrier. The results showed that patient serum with high angiopoietin-2 concentrations increased (and addition of angiopoitein-1 reversed) gap formation of endothelial test cells.
The in vivo relevance of this finding was documented by measuring the escape of intravenously-administered Evans Blue out of the pulmonary vascular bed in animals pretretated with angiopoietin-2; enhanced leakage was found, though the mice did not show the signs of sickness that one typically encounters in sepsis.
Finally, the authors addressed the molecular mechanisms involved in the angiopoietin-2–induced endothelial cell hyperpermeability.
It has been known for some time that myosin-driven (21) cell contractions caused by myosin light chain phosphorylation (22) via myosin light chain kinase (23) underlie endothelial cell gap formation (24). The authors showed that, in their test system, angiopoietin-2 dose-dependently caused myosin light chain phosphorylation. This was inhibited by angiopoietin-1. Upregulation of myosin light chain phosphorylation is known to require activation of the small GTPase RhoA. It was therefore not surprising that in this preparation the specific inhibitor of Rho-kinase Y27632 completely abolished angiopoietin-2–induced myosin light chain phosphorylation.
That the effects of angiopoietin-2 are mediated via inactivation of the Tie2 receptor was finally proven by a knock-down experiment. The result suggested that the Rho-kinase/myosin light chain phosphorylation cascade is triggered when the antagonistic angiopoietin-2 abrogates the phosphorylation of the Tie-2 receptor, which is normally caused by the agonist angiopoietin-1.
These findings strongly suggest a causal role of angiopoietin-2 in the genesis of the pulmonary vascular leak of sepsis. Whether the effects of angiopoietin-2 on the pulmonary microcirculation are the only mechanism or whether additional upstream actions on larger vessels contribute requires further study.
Of course, it is too early to tell whether these impressive mechanistic insights will ultimately translate into novel interventions, but the appeal of modulating the above cascade would be that instead of relatively broadly acting, unfocused immunomodulatory and anti-inflammatory interventions fraught with an array of potential side effects, the angiopoietin-2–based strategy would be narrowly focused. Unless unforeseen effects show up, one would anticipate fewer side effects. Of course, confirmation of these stimulating findings and controlled clinical trials are required.
|
 |
References |
|---|
 Top References References References |
|---|
It had been known for considerable time that because of the propagation of the pressure wave in the elastic central conduit arteries, the instantaneous pressure is not uniform throughout the central conduit arteries (1). It has only been relatively recently, however, that the potential clinical relevance of this phenomenon has been recognized and that adequate methodology has been developed to further look into this problem (2,3). One spin-off of this debate has been the recognition that blood pressure (BP) measured in the brachial artery may differ to a variable degree from the pressure existing in the aorta, that this pressure difference is affected by age and disease states and that antihypertensive agents differ with respect to their impact on the BP difference between brachial and central aortic pressures. Despite the remarkable findings in some smaller studies (4–6), what had been lacking so far was well-documented evidence of this phenomenon in a sufficiently large cohort and proof that the results impact on clinical outcomes.
The ASCOT (Anglo-Scandinavian Cardiac Outcomes Trial) study was designed to compare "conventional" antihypertensive agents, i.e., 
blockers and diuretics (Atenolol±thiazide-based diuretics) with more "modern" antihypertensive drugs, i.e., a calcium channel blocker and ACE inhibitor (Amlodipin±Perindopril). The outcome of the primary study had been published before (7), which documented the superiority of the Amlodipin/Enalapril combination on fatal and nonfatal stroke, total cardiovascular events, and procedures, as well as on all-cause mortality, although the difference was not significant for the primary endpoint of nonfatal myocardial infarction and fatal chronic heart disease. In an effort to further elucidate potential mechanisms accounting for any differences between the two antihypertensive strategies despite no differences in BP as measured conventionally at the level of the brachial artery ("beyond BP"), Bryan Williams and colleagues examined in a preplanned substudy (Conduit Artery Function Evaluation [CAFE] study) whether the two antihypertensive strategies differed with respect to their effect on central pressure relative to peripheral arterial BP (8). To this end, out of the total of 19,257 patients in the ASCOT study in 5 study centers, 2199 patients with systolic BP >160 mmHg, diastolic BP >100 mmHg, or treated BP >140 mmHg systolic or >90 mmHg diastolic, and with additional cardiovascular risk factors were entered into the CAFE substudy and followed for 3 years. The study had a PROBE design (prospective, randomized, open, blinded end point evaluation). The brachial BP was measured with the semiautomated oscillometric Omron 705CP device. The calculation of the derived central pressure was based on measurements using radial artery applanation tonometry (SphygmoCor) and calculation with a dedicated software applying a validated transfer function. In addition, pulse wave analysis was performed in a subset of patients.
A comment concerning the rationale behind the study may be appropriate. The central aortic pressure (and by implication the pressure determining coronary and renal perfusion) is not simply given as the ratio between cardiac output and peripheral vascular resistance according to the law of Ohm. Rather, the time-dependent pressure wave is modified by the elastic properties of the central arteries and, in addition, by the backreflection of the outgoing wave in the periphery, which is then superimposed on the outgoing central pulse wave in late systole and can be quantitated by calculating the augmentation index (5). In young individuals with no structural or functional changes of the central arteries, the difference between peripheral and central BP is negligible, whereas in elderly subjects or individuals with changes in the structure and function of central vessel walls, the deviation may be substantial and clinically relevant (6,9). Specifically in renal patients, Gerard London showed marked deviations of the pulse profile from normal with exaggerated systolic upstroke secondary to increased stiffness of the vessel wall, and, during late systole, premature return of the pulse wave reflected in the periphery (10).
What findings were obtained in the CAFE substudy? At the end of the study the brachial artery pressure values were similar in the two groups on Atenolol±thiazide-based diuretics and on Amlodipin±Perindopril, respectively. The measured radial and the calculated central pulse wave forms were strikingly different, however; in patients treated with Atenolol, the radial pulse wave form was broader with a more prominent late systolic peak. The derived central aortic systolic pressure was substantially lower (by 4.3 mmHg, 3.3 to 5.4, 95% confidence interval) in the group on Amlodipin±Perindopril compared with the group on Atenolol±thiazide-based diuretics. The difference was significant and consistent during repeated measurements with time; there were small differences in central aortic diastolic pressure values as well.
As a potential explanation for the higher central aortic pressure in the patients on Atenolol±thiazide-based diuretics compared to the central aortic pressure in patients on Amlodipin±Perindopril, the authors could exclude an increase in the amplitude of the outgoing systolic pressure wave, but a plausible explanation is the increased augmentation of the systolic pressure wave in the aorta; the percent increase in the central systolic pressure wave attributable to wave reflection (augmentation index) was 6.5%.
These findings are not an innocent academic exercise, but impact the cardiovascular and renal outcomes in the CAFE study. Using the Cox proportional hazards technique in the different Cox models, each of the following were significantly associated with the composite end point comprising all cardiovascular events and cardiovascular procedures as well as development of renal impairment: calculated central aortic pulse pressure, central aortic pressure wave augmentation, as well as outgoing pressure wave height and brachial pulse pressure (as a crude index of aortic stiffness). Even after adjustment for several baseline confounders, this relation continued to be significant, particularly for central aortic pulse pressure.
Why should Atenolol±thiazide-based diuretics cause such impressive changes in wave reflection? The earlier return of the pulse wave is not explained by differences in arterial pulse wave velocity, because the measured pulse wave velocity between the groups on Atenolol±thiazide-based diuretics compared with the group on Amlodipin±Perindopril was not significantly different. Another possibility to consider is that, in the group on Atenolol±thiazide-based diuretics, the outgoing wave is reflected closer to the aorta, thus explaining the earlier return of the wave potentially as a result of arterial vasoconstriction (11) or remodeling (12), or conversely, vasodilatation and remodeling by calcium channel blockers. This interpretation would be in line with the results obtained in the Preterax in Regression of Arterial Stiffness in a Controlled Double-Blind study (REASON) study (13). A final possibility would be that, in patients on blockers, bradycardia prolonged the ejection time so that the incoming reflected wave was more extensively superimposed upon the outgoing wave (14).
The above documented behavior of central pressure and pulse wave contours on blocker treatment may well explain why in previous studies blockers—despite similar achieved BP—were inferior to alternative antihypertensive treatments concerning reduction of left ventricular hypertrophy (15), of carotid intima-media thickness (16), as well as modification of resistance artery structure (17–19).
Because of the marked abnormalities of the structure and elasticity of central arteries in patients with renal dysfunction (20,21), even at relatively early stages of reduced GFR (22), the above findings are of considerable interest to the nephrologist. Indeed, it has been shown that indices of elasticity of central arteries respond differently to different antihypertensive agents (6,23), and BP-independent effects of ACE inhibitors on aortic elasticity have been shown in renal patients as well (24).
Because not only perfusion of the coronary vessels, but also of the vessels of the kidney, is driven by the central—and not by the peripheral—arterial pressure, it is conceivable that some of the superior renoprotection of antihypertensive agents such as ACE inhibitors and angiotensin receptor blockers (25,26) is accounted for, at least in part, by their beneficial effect on central pressure. In the ongoing discussion of whether ACE inhibitors or angiotensin receptor blockers provide renoprotection beyond BP (25,27), it is definitely relevant to consider also the effects of blockade of the renin-angiotensin system on central pressure.
|
References |
|---|
|
TopReferencesReferences References |
|---|
Ever since the pioneer studies of Abraham Trembley in 1740, who had documented that "a tiny piece of an older hydra could grow into the entire animal just the way the embryo did" (1), the amazing potential of multipotent cells in lower species to regenerate organs and body parts provoked great interest. In the recent past, the recognition that even in higher species circulating progenitor cells are capable of amazing organ repair (2) has led to a scientific revolution with fast-moving frontiers and intense efforts to translate basic findings to the clinical setting.
One area where great hopes exist in the therapeutic potential of stem cells is the treatment of acute myocardial infarction (3,4). In clinical studies, bone marrow–derived cells were either administered via intracoronary infusion or were mobilized from the bone marrow via administration of cytokines such as granulocyte colony stimulating factor (G-CSF) or stem cell factor (5).
After early reperfusion strategies have markedly lowered the early mortality of myocardial infarction, the major challenge for stem cell therapy is the prevention or attenuation of postinfarction heart failure.
As it so happens, two recent communications reported long-term outcomes in sizable cohorts of patients. In the first study, after acute myocardial infarction patients received either autologous bone marrow cells infused into the infarct-related artery by stop-flow balloon catheter technique or placebo (6)—the so-called BOOST trial (Bone Marrow Transfer to Enhance ST Elevation Infarct Regeneration). The other study assessed the late outcome in patients with acute myocardial infarction in whom stem cells had been mobilized by administration of G-CSF (7); with a grain of excess optimism the trial was called REVIVAL-2 study (Regenerate Vital Myocardium by Vigorous Activation of Bone Marrow Stem Cells).
The long-term results of either study are somewhat sobering. In agreement with several small, less well-controlled trials or trials not specifically designed for this purpose (8), the initial results of the BOOST study had been promising indeed (6): within 6 months a significant improvement of global as well as regional left ventricular ejection fractions (LVEF) had been noted (6). This was accompanied by evidence of improved diastolic function (9). There had been no major side effects. The 18-month long-term results, however, which are presented now, are not earth-shaking: the difference in LVEF, which had been significant at 6 months, was no longer significant after 18 months, although the speed of recovery had obviously been faster (10).
How about the competing procedure, stem cell mobilization by G-CSF in patients with ST-segment elevation acute myocardial infarction (STEMI)? G-CSF was administered with the rationale that stem cells mobilized by the cytokine home in to the site of myocardial injury. Previous small studies had suggested benefit (11,12); they can all be criticized, however, for one reason or another, e.g., problems with control patients, randomization, blinding, sample size, etc. (13). The hypothesis that mobilizing bone marrow–derived precursors provides a clinically relevant benefit was subjected to a rigorous test in an adequately-sized randomized study, comprising more patients than all preceding studies taken together. The trial included patients with STEMI who had been successfully reperfused within 12 h after onset of symptoms. They received either placebo or 10 µg/kg G-CSF for 5 d. The primary end point was reduction of left ventricular (LV) infarct size by Tc99m sestamibi scintigraphy at baseline and after 4 as well as 6 months. Secondary end points were, among others, improvement of LVEF (by magnetic resonance imaging) and restenosis (by angiography).
Not surprisingly, the patients receiving G-CSF experienced an increase in CD34+ cells (marker of bone marrow–derived precursors) and white blood cells, a decrease in platelets, and an increase in C-reactive protein, lactate dehydrogenase, and alkaline phosphatase. They experienced slightly more general side effects than the placebo group.
There were no differences of infarct size between the groups at baseline or at follow-up after 4 and 6 months. LV function, evaluated as LVEF using magnetic resonance imaging, and the values of the LV end-systolic volume index were absolutely identical between the groups. Finally, angiography showed no difference of restenosis rate between the groups, at least in patients with bare (i.e., non–drug-coated) stents. In contrast with previous reports (11), there was no evidence of excess restenosis. Thus, the good news is that there were no serious side effects. The bad news is that there was no demonstrable benefit either.
Several possibilities may account for the negative outcome. Acute myocardial infarction releases progenitor cells from the bone marrow by upregulating the secretion of cytokines (14). It is possible, therefore, that the functional consequences of a further increase caused by exogenous cytokines is limited. After mobilization from the bone marrow, the progenitor cells may also have not been able to properly home to the damaged cardiac tissue (this shortcoming is circumvented by intracoronary administration) (6).
A further possibility is that the cytokine took too much time to mobilize stem cells: in experimental studies homing was seen only in the first days immediately after myocardial infarction (15–18). That the time window may be an important limitation is also suggested by the first observations in the Reinfusion of Enriched Progenitor Cells and Infarct Remodeling in Acute Myocardial Infarction (REPAIR-AMI) trial (4) that improvement in LV function is correlated to the time interval between acute event and intracoronary stem cell administration. If this is the explanation, intracoronary administration would obviously be the way to go.
Is this negative study outcome the requiem for stem cell therapy in general? We hope not. The authors of this negative study did an excellent job in recruiting an adequately-sized cohort of randomized patients and in applying state-of-the art methodology to evaluate heart function. It is plausible that past investigators might have been fooled by the spontaneous improvement of LV function, which is well documented in the placebo group of Zohlnhöfer et al. Factors such as reversal of myocardial stunning, removal of necrotic tissue, edema or infiltrating cells, and finally contraction of collagen may change the infarct tissue in such a fashion that this is easily confounded with treatment success (13).
One reason why it has become so difficult to improve outcomes is that modern early-reperfusion techniques have reduced early mortality so much that it is difficult to do better. After the acute window is closed, the main challenge is to improve the long-term outcome over and above today’s success rates with blockade of the renin-angiotensin system, statins, and antiplatelet regimens—not an easy task.
The initial, perhaps naïve, idea of how stem cells operate was direct transdifferentiation of bone marrow–derived precursors into mature cardiomyocytes (19), an idea which is not supported by experimental data (20,21) A more likely explanation, well known to nephrologists from what happens in the repair phase of acute renal failure (22), is the idea that stem cells act by paracrine secretions, thus creating a proangiogenic, antiapoptotic, and proproliferative environment (15,23,24). Bone marrow–derived mononuclear cells were, for instance, shown to express VEGF, basic fibroblast growth factor, and angiopoietins (25–27). Another current unknown is whether it is possible to modulate the (limited) number of multipotent resident cardiac stem cells (28).
Although the results of the G-CSF stem cell mobilization study are outright negative, and the long-term outcome of the BOOST trial—despite initially positive results—are sobering, it would be too early to count the stem cell approach out. But after these studies the task will be much tougher.
|
References |
|---|
|
TopReferencesReferences References |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
HOME
CURRENT ISSUE
ARCHIVES
JASN Express
ONLINE SUBMISSION
AUTHOR INFO
EDITORIAL BOARD SUBSCRIBE FEEDBACK ALERTS HELP |
