Hypoxia-inducible factors and the prevention of acute organ injury

Hypoxic preconditioning has long been considered as organ-protective, and its clinical usage has been suggested in elective procedures, such as coronary surgery and organ transplantation. Although the mechanisms have not been clearly elucidated, it has been postulated that changes in cell-membrane composition and upregulation of various cellular protective mechanisms are responsible for a better tolerance of acute injury. Remote preconditioning (i.e., hypoxic stress in one organ conferring resistance to acute hypoxia in other organs) suggests organ cross-talk, perhaps mediated by cytokines and the immune system.


Introduction
Hypoxic preconditioning has long been considered as organ-protective, and its clinical usage has been suggested in elective procedures, such as coronary surgery and organ transplantation. Although the mechanisms have not been clearly elucidated, it has been postulated that changes in cell-membrane composition and upregulation of various cellular protective mechanisms are responsible for a better tolerance of acute injury. Remote preconditioning (i.e., hypoxic stress in one organ conferring resistance to acute hypoxia in other organs) suggests organ cross-talk, perhaps mediated by cytokines and the immune system.
Increased expression of heme-oxygenase (HO)-1, heatshock proteins (HSP), growth factors such as vascular endothelial factor (VEGF), and erythropoietin (EPO) are among the numerous adaptive responses to sublethal injury that are believed to participate in tissue tolerance during subsequent stress. EPO, for instance, is a ubiquitous pleiotropic survival and growth factor that attenuates experimental acute injury in various organ systems, including neuronal, retinal, cardiac, renal, and hepatic tissues. Its clinical effi cacy, though suggested in critically ill patients, is yet to be defi ned [1].
Th e expression of these protective mediators and many others is regulated by hypoxia-sensing mechanisms through the induction and stabilization of so called hypoxia-inducible factors (HIF) [2]. In this chapter, we will outline the control and action of HIF as key regulators of hypoxic adaptive response, and particularly examine HIF expression during hypoxic stress. We shall discuss recently developed measures that enable HIF signal modifi cation and describe their potential use in conferring tissue tolerance during incipient organ injury.

HIF regulation and action
HIFs are heterodimers ( Fig. 1), composed of a constitutive β-subunit (HIF-β) and one of three diff erent oxygendependent and transcriptionally active α-subunits, among which HIF-1α and -2α are acknowledged as promo tors of hypoxia adaptation, whereas the role of HIF-3α remains unclear. Under normoxia, HIF-α subunits are constantly produced, but not allowed to accu mulate, since they are rapidly hydroxylated by oxygen-dependent HIF prolyl-4hydroxylase domain enzymes (PHD), subsequently captured by the ubiquitin ligase Von-Hippel-Lindau protein (VHL), and degraded by the proteasome. Under oxygen defi ciency, PHD activity is reduced, HIF-α accumulates within the cytosol, αβ-dimers are formed, translocate into the nucleus, and bind to hypoxia response elements (HREs) in the promoter enhancer region of genes, which are subsequently transactivated [2][3][4].
Th e biological eff ects of the more than 100 acknowledged HIF target genes are multiple, and include key steps in cell metabolism and survival. Many of the HIFtarget genes constitute a reasonable adaptation to hypoxia, such as erythropoiesis (EPO), increased glucose uptake (glucose transporter-1), switch of metabolism to glycolysis (several key enzymes of glycolysis), increased lactate utilization (lactate dehydrogenase), angiogenesis (VEGF), vasodilation (inducible nitric oxide synthase [iNOS]), removal of protons (carbonic anhydrase 9), and scavenging of free radicals (HO-1) [2][3][4].

Biological and rherapeutic modes of HIF activation
Every cell type has the potential to upregulate HIF, principally by the inhibition of PHD, under conditions when cellular oxygen demand exceeds oxygen supply, Hypoxia-inducible factors and the prevention of acute organ injury Samuel N Heyman 1 *, Seymour Rosen 2 , Christian Rosenberger 3 This article is one of eleven reviews selected from the Annual Update in Intensive Care and Emergency Medicine 2011 (Springer Verlag) and co-published as a series in Critical Care. Other articles in the series can be found online at http://ccforum.com/series/annual. Further information about the Annual Update in Intensive Care and Emergency Medicine is available from http://www.springer.com/series/8901 namely under cellular hypoxia. However, the threshold and extent of HIF activation may depend on the hypoxic stimulus and cell type involved. To some extent, these cellular variations may refl ect diff erent expression of various PHD isoforms in diff erent tissues [5][6][7].
As HIF stimulation may potentiate hypoxia tolerance, studies were conducted to explore its clinical application. Widespread experimental hypoxic stimuli are listed in Table 1, all acting principally by the control of HIF-α degra dation, initiated by PHDs. Except for carbon monoxide exposure, which is currently being tested in patients, none of these stimuli seems suitable for preconditional HIF activation in humans.
Apart from hypoxic stabilization, widely proven in vivo, HIF activation has also been demonstrated to occur under normal ambient oxygen tensions, mostly in cell cultures challenged with cytokines and growth factors. However, under stress, oxygen demand likely is increased, thus possibly leading to intracellular hypoxia even in cells kept under room air. For technical reasons, it is probably impossible to rule out such local cellular hypoxia that may exist predominantly within the mitochondria. Beyond this academic distinction between true cellular hypoxia and normoxia, it is important to recognize that clinical conditions, like infl ammation, infection and sepsis, may lead to HIF activation. Th us, theoretically, cytokines or growth factors could be used for preconditional HIF activation in humans.
Although not a reasonable therapeutic intervention, strong and stable normoxic HIF activation can be Hypoxia inhibits PHDs and leads to HIF-α accumulation with HIF-β, and the αβ heterodimer translocates into the nucleus, binds with hypoxia-response elements (HRE) and activates numerous genes important in cell metabolism, proliferation and survival. Many of these genes play a central role in injury tolerance and promotion of tissue oxygenation, such as erythropoietin (EPO), vascular endothelial growth factor (VEGF), inducible NO synthase (iNOS), heme oxygenase (HO)-1, glucose transporter-1, or carbonic anhydrase (CA)-9. Underscored is the inactivation of the HIF-HRE axis by hypoxia, which can be mimicked by carbon monoxide (functional anemia) or by transition metals like cobaltous chloride. Hypoxia-mimetic PHD inhibitors (PHD-I) are potent newly developed measures in the induction of the HIF-HRE axis. For simplicity, numerous additional factors involved in HIF regulation and action are not included in this cartoon and the reader is referred to comprehensive reviews such as references [3,12].  [8]. Additional experimental probes for enhancing HIF signal are by transfection with PHD siRNA [9] or with the generation of constitutively active HIF-α transgenes [10]. So-called hypoxia mimetics block PHD activity, thus upregulating HIF under normoxia. PHDs require 2-oxoglutarate and ferrous iron as co-substrates. Nonspecifi c PHD inhibitors are either 2-oxoglutarate analogues or interfere with Fe 2+ . Recently, more specifi c PHD inhibitors (PHD) have been synthesized [11], and are currently being tested in animal and human studies. Figure 1 represents a simplistic scheme of the canonical HIF regulation and action. Recent discoveries underscore a host of additional compound biological pathways, associated with the regulation of the HIF signal, including the control of HIF synthesis, HIF controlling PHD synthesis, putative competing/intervening impacts of HIF-3α and PHD-3, cross-talk of HIF and other key regulators of gene expression (STAT, p-300 and others), further modifi cation of HIF-α activity at the level of DNA hypoxiaresponsive elements by small ubiquitin-like modifi ers (SUMO) and factor inhibiting HIF (FIH), and the eff ect of reactive oxygen species (ROS), NO and Krebs cycle metabolites on HIF degradation. Th ese complex pathways are beyond the scope of this review, and the interested reader is referred to additional references [3,5,[12][13][14][15][16][17][18].

HIF expression under hypoxic stress and tissue injury
Th e kidney serves as an excellent example for understanding HIF expression under hypoxic stress. Renal oxygenation is very heterogeneous, with PO 2 falling to levels as low as 25 mmHg in the outer medulla under normal physiologic conditions and to even lower values in the papilla [3,4,19]. Changes in renal parenchymal microcirculation and oxygenation have been thoroughly investigated in acute and chronic renal disorders [19,20]. Finally, the complex renal anatomy in which diff erent cell types are in close proximity to regions with comparable ambient oxygenation, enables comparisons of cellular HIF response.
Interestingly, HIF expression is below detection threshold by immunostaining in the renal medulla, despite low physiologic ambient oxygenation (It should be emphasized that this statement regarding negative HIF immuno staining in the normally hypoxic medulla relates to kidneys perfusion-fi xed in vivo without an interruption of renal oxygenation before fi xation. Other modes of tissue harvesting for HIF determination, either by immunostaining or by molecular biology techniques may be falsely positive, as hypoxia-induced inhibition of PHD activity is instantaneous, and may lead to HIF-α stabilization even over short periods of hypoxia). Conceivably, this refl ects the plasticity of HIF control to adjust for 'physiologically normal' oxygenation (i.e., adjusted rates of HIF-α generation and degradation under normal conditions. Enhanced renal HIF-α is noted in rodents subjected to hypoxia or to inhaled carbon monoxide (chemical hypoxia) [21], and in hypoxic isolated perfused kidneys [22]. Diff erent cells express diverse HIF isoforms: Where as tubular segments express HIF-1α, HIF-2α is principally produced by vascular endothelial and interstitial cells [21][22][23]. Interestingly, HIF-dependent genes are also selectively expressed in diff erent cell types. For instance HIF-2-triggered EPO generation is specifi cally found in interstitial cells in the deep cortex [24]. In hypoxic isolated perfused kidneys, attenuation of severe medullary hypoxia by the inhibition of tubular transport markedly enhanced HIF expression, probably underscoring a window of opportunity to generate HIF and HIF-mediated adaptive responses only under moderate and sublethal hypoxic stress [22]. Th is pattern is consistent with HIF expression at the border of renal infarct zones only, indicating that dying cells within the critically ischemic region are incapable of mounting a hypoxia adaptive response [25]. We also found that HIF-α isoforms are stabilized in acute hypoxic stress, predominantly in the cortex in rhabdomyolysis-induced kidney injury [26], in the outer stripe of the outer medulla following ischemia and reperfusion [27,28], or in the inner stripe and inner medulla following the induction of distal tubular hypoxic injury by radiocontrast agent, or after the inhibition of prostaglandin or NO synthesis or with their combinations [23]. Outer medullary HIF stabilization is also noted in chronic tubulointerstitial disease [29] and in experi mental diabetes [30], again spatially distributed in areas with proven hypoxia. HIF was also detected in biopsies from transplanted kidneys [31]. Th us, HIF immunostaining is chronologically and spatially distributed in renal regions with abnormally low PO 2 .
Normal mice subjected to warm ischemia and reperfusion display limited injury only, as compared with extensive damage in HIF (+/-) mice [32]. Th us, the impor tance of mounting an HIF response during hypoxic stress is undeniable.
Hypoxia-driven HIF stabilization during hypoxic stress has been encountered in other organs as well. HIF-1α and PHD-2 expression increased in the neonatal rat brain following hypoxia [33] and HIF was detected in the hypoxic subendocardium [34] and in the ischemic liver [27]. HIF is also found within hypoxic regions in tumors, and may play an important role in tumor progression via upregulation of growth promoting and angiogenic factors [35].

Potential usage of HIF modulation in clinical practice
Th e impact of HIF stimulation on the expression of HIFdependent tissue-protective genes led to the expectation that timely upstream HIF stimulation may have great potential in the protection of endangered organs by down stream induction of protective genes [12]. Indeed, repeated systemic hypoxia, for instance, results in enhanced expression of renal HIF and HIF-dependent genes and attenuates warm-ischemic injury [36].
Th e use of hypoxia-mimetic PHD inhibitors is a promis ing potential new treatment option in diseases such as myocardial infarction, stroke, renal or liver injury, peripheral vascular disease, or severe anemia. Studies with PHD inhibitors and other manipulations of HIF upregulation favor this hypothesis [11].

Anemia
Specifi c PHD inhibitors induce HIF-2α expression in interstitial fi broblasts in the deep cortex [24], enhance erythropoietin generation, and were found to provoke erythrocytosis in primates [37]. Phase 2 clinical trials in patients with chronic kidney disease are currently under way, studying the eff ect of oral PHD inhibitors as potential substitutes to EPO injection.

Acute kidney injury
Th e potential protective impact of HIF upregulation by PHD inhibitors has been extensively studied in acute kidney injury. In isolated kidneys perfused with lowoxygen containing medium, pre-treatment with a PHD inhibitor improved renal blood fl ow and attenuated medullary hypoxic damage [38]. Conditional inactivation of VHL in mice (hence HIF stabilization) resulted in tolerance to renal ischemia and reperfusion [8] and to rhabdomyolysis-induced acute kidney injury (Rosenberger C, unpublished data). Whereas gene trans fer of negative-dominant HIF led to severe damage in the normally hypoxic renal medulla in intact rats, transfer of constitutively active HIF (HIF/VP16) induced expression of various HIF-regulated genes and protected the medulla against acute ischemic insults [39]. Furthermore, in rats and mice subjected to warm ischemia and refl ow, PHD inhibitors and carbon monoxide pre-treatment (i.e., functional anemia) markedly attenuated kidney damage and dysfunction [32,40]. Donor pre-treatment with a PHD inhibitor also prevented graft injury and prolonged survival in an allogenic kidney transplant model in rats [41]. Finally, rats preconditioned by carbon monoxide, displayed reduced cisplatin renal toxicity, with attenuation of renal dysfunction and the extent of tubular apoptosis and necrosis [42]. Taken together, all these observations indicate that HIF stabilization seemingly is a promising novel interventional strategy in acute kidney injuries [12].

Myocardial injury
Activation of the HIF system has also been found to be cardioprotective. In a model of myocardial ischemia in rabbits, pre-treatment with a PHD inhibitor induced robust expression of HO-1 and markedly attenuated infarct size and myocardial infl ammation [43]. In another report, PHD inhibitors did not reduce infarct size, but improved left ventricular function and prevented remodeling [44]. In the same fashion, selective silencing of PHD-2 with siRNA 24 h before global myocardial ischemia/reperfusion in mice reduced the infarct size by 70% and markedly improved left ventricular systolic function [9]. Remote preconditioning by intermittent renal artery occlusion also resulted in cardiac protection, conceivably through PHD inhibition [45].
Enhanced levels of PHD-3 were traced in the hibernating myocardium [34] and in end-stage heart failure in humans, associated also with elevated HIF-3α [46] (which may act as a competitive inhibitor of active HIF-α isoforms [14]). Th us, PHD inhibitors may conceivably also be benefi cial in these disorders. Finally, cardioprotection during heat acclimation is also mediated in part by HIF upregulation [47], providing another potential situation for the administration of PHD inhibitors.

Neuronal injuries
Th e eff ect of PHD inhibitors has also been assessed in disorders of the central nervous system. In vitro, rotenoneinduced neuronal apoptosis was attenuated and autophagy increased, as the result of enhanced HIF following deferoxamine administration [48]. In vivo, PHD inhibitors have shown promising results in the attenuation of ischemic stroke [49], and might be neuroprotective in metabolic chronic neurodegenerative conditions [50]. However, studies showing inhibition of PHD-1 by ROS suggest non-HIF-mediated neuronal protection under normoxic conditions [51].

Lung injury
Preterm lambs developing respiratory distress syndrome display upregulation of PHDs with a reciprocal fall in HIF-α isoforms and HIF-dependent VEGF [53]. Th is observation implies that PHD inhibitors might have therapeutic potential in this clinical setup.

Liver disease
Hepatic HIF-1α is upregulated following warm ischemia [27], and is required for restoration of gluconeogenesis in the regenerating liver [52], implying yet another potential use for PHD inhibitors in acute liver disease.

Peripheral vascular disease
In a model of limb ischemia in mice, PHD inhibitors enhanced HIF expression and downstream VEGF and VEGF-receptor Flk-1, leading to improved capillary density, indicating a potential therapeutic use of PHD inhibi tors in promoting angiogenesis in ischemic diseases, such as severe peripheral vascular disease [54]. Transfection with HIF-1α, combined with PHD inhibitortreated bone marrow-derived angiogenic cells increased perfusion, motor function, and limb salvage in old mice with ischemic hind limbs [55]. Results of a phase-1 study in patients with critical limb ischemia indicate that transfection with a constitutively active form of HIF-1α might also promote limb salvage [10]. Further clinical trials with PHD inhibitors are currently under way in burn wound healing and salvage of critically ischemic limbs.

Oxidative stress
Enhanced cellular ROS concentrations, as happens with shock and tissue hypoxia, result in increased PHD activity, and this eff ect is antagonized by ROS scavengers [15]. Th is situation may lead to HIF de-stabilization and inadequate HIF response to hypoxia. For example, hypoxia-mediated HIF expression in the diabetic renal medulla is substantially improved by the administration of the membrane-permeable superoxide dismutase mimetic tempol [30]. It is, therefore, tempting to assume that ROS scavengers, as well as PHD inhibitors may improve tissue adaptive responses to hypoxia, coupled with oxidative stress. However, contradicting evidence exists, indicating that ROS might trigger HIF in the absence of hypoxia. Th is has been suggested by studying liver tissue in acetaminophen-induced liver injury, before the development of overt liver injury and hypoxia [56], and in aged well-fed animals [57]. Th e role of HIF stimulation during oxidative stress therefore needs further assessment.

Important considerations
HIF stimulation is not all-protective. Th e wide range of HIF-dependent genes, and its tight cross-communication with other key regulators of gene expression [13,58,59] raise concern regarding concomitant non-selective activation of protective as well as harmful systems. Among potential unwanted outcomes is the enhancement of tumor growth [60], promotion of fi brosis [61] or the induction of pre-eclampsia in pregnant women [62]. Indeed, whereas HIF activation is considered renoprotec tive in acute kidney injury, it may play a role in the progression of chronic kidney disease and certainly is an important factor in the promotion of renal malignancy [3,20].
Diverse characteristics and distribution patterns of diff erent PHDs [5][6][7] and particular actions of various PHD inhibitors [11,37] might enable selective manipulation of the HIF system in a more desired way, selectively favoring advantageous HIF-dependent responses in preferred tissues. Furthermore, it is believed that activation of adverse responses requires protracted HIF stimulation, whereas short-term and transient HIF activa tion might suffi ce to activate tissue-protective systems without continuing induction of harmful systems. However, this concept needs confi rmation in clinical trials.

Conclusion
Elucidating the mechanisms involved in HIF-mediated cellular responses to acute hypoxic stress has led to the discovery of novel potential therapeutic options for the prevention or attenuation of tissue injury. Th e non-selective enhancement of gene expression by current modes of HIF augmentation warrants caution, since undesired enhancement of certain genes may be hazardous.
We anticipate that in the coming years the use of PHD inhibitors and other stimulants of the HIF system will be tested in many clinical scenarios associated with critical care and emergency medicine, while HIF silencing strategies may be tested in chronic diseases, such as malignancies and disorders with enhanced tissue scarring.