Sensing and Responding to Hypoxia, Molecular and Physiological Mechanisms1 (2024)

,

Abstract

WHAT IS NORMOXIA FOR THE MITOCHONDRION?

Metazoan life has evolved to take advantage of the high energy yield that can be derived from oxygen-based respiration and in all eucaryotes aerobic production of ATP by oxidative phosphorylation takes place in the mitochondria whereby each mole of glucose is oxidized by 6 moles of O₂ to yield, theoretically, 36 moles of ATP. This process is 18 times more efficient than anaerobic glycolysis which only yields 2 moles of ATP per mole of glucose. When mitochondria formed an endosymbiotic association with cells about 3,000 million years ago, atmospheric oxygen levels were probably less than 0.1% of present values (Gray et al., 1999). Today, in the intracellular microenvironment of the cell mitocondria continue to experience similar low oxygen conditions, and it is very likely that they are adapted to, or dependant on, such conditions. For example the P₅₀ for PO₂ dependant mitochondrial oxygen consumption is less than 1.0 mmHg, (Gnaigner et al., 2000). Indeed it appears that increasing oxygen levels amplifies uncoupled proton leak in mitochondria and handicaps their bioenergetic efficiency such that at air saturation oxygen conditions (the conditions for almost all isolated mitochondrial studies) mitochondria are under considerable oxidative stress (Gnaiger et al., 2000).

Because oxygen plays such a critical role in metabolism most animals are extremely sensitive to changes in ambient oxygen levels and a reduction in normoxic oxygen supply offers a serious challenge. On encountering hypoxia the first response is to enhance oxygen delivery to the deprived tissue. If this is insufficient alternate non oxygen dependent ATP producing pathways are recruited and as a last resort mechanisms to reduce oxygen demand are activated. The ability to detect oxygen deprivation is, therefore, critical to the survival of all forms of aerobic life (Bunn and Poynton, 1996).

HYPOXIA SENSING

Biological signals for detecting oxygen insufficiency occurs at three levels, first systemic, then cellular/genetic and finally energy crisis.

Systemic

Hypoxia is first sensed by central and peripheral chemoreceptors. Little is known of their identity; arterial oxygen receptors may involve membrane located oxygen sensitive K⁺ channels (Prabhakar and Overholt, 2000), the carotid body may have mitochondrion oxygen sensors acting via cytochrome aa3 redox levels (Lahiri et al., 1995). On detecting hypoxia the central and peripheral oxygen sensors immediately activate mechanisms to enhance oxygen delivery. These physiological responses are well understood (see review by Lutz and Storey, 1997). They function to produce an immediate increase in oxygen uptake by increasing lung ventilation and enhancing oxygen extraction efficiency. Correspondingly, oxygen delivery is improved through a greater cardiac output and an increase in tissue perfusion.

Cellular/genetic

Changes in oxygen availability can stimulate the activity of hypoxia-sensitive genes which elicit a range of short and long term adaptations to enhance oxygen uptake and delivery. But in order to respond hypoxia must initially be detected. The search for the oxygen sensor(s) is a rich and rapidly growing field and a wide variety of molecules have been proposed.

Oxygen sensors

There is substantial evidence that membrane proteins containing heme groups may be oxygen sensors (Bunn and Poynton, 1996). These have been identified as cell membrane bound b-cytochrome coupled proteins, which are universal in distribution, from bacteria to mammals (Bunn and Poynton, 1996). In particular the cytochrome b558/NADPH oxidase complex is thought to be an oxygen sensor in a range of cell types including phagocytes, oxygen sensing cells of the carotid body and airway chemoreceptor cells.

Cytoplasmic redox-sensitive sites such as NADPH oxidases have also been suggested (Prabhakar and Overholt, 2000). In mammalian systems hypoxia has been shown to interfere with an iron dependent interaction between the hypoxia responsive transcription factor HIF-1 alpha and its partner VHL (Maxwell et al., 1999). In the normoxia, VHL targets cytoplasmic HIF-1 alpha for proteasomal degradation. In hypoxia, VHL no longer associates with HIF-1 alpha and HIF-1 alpha translocates to the nucleus, where HIF-1 alpha participates in activation of hypoxia inducible genes (Maxwell et al., 1999).

Chandel et al. (1998) have proposed that the mitochondrion may act as the oxygen sensor. They found that hypoxic induction of oxygen regulated genes was blocked in hepatocyte rho-0 cells which lack electron transport chain activity. Cytochrome aa3 has been identified as the mitochondrial oxygen sensor. In hypoxia reduction of O₂ to H₂O by cytochrome aa3 (complex IV) is inhibited, causing a release of electrons upstream at complex III and generation of O₂⁻. The mitochondrial model of cytochrome aa3 as the oxygen sensor is consistent with an increase in cellular ROS levels in hypoxia (Wenger, 2000).

Hypoxia induced transcription factors

The most well described hypoxia induced transcription factor, hypoxia inducible factor-1 (HIF-1) activates a battery of hypoxia responsive genes such as vascular endothelial growth factor (Forsythe et al., 1996), phosphofruvtokinase 1 (sem*nza et al., 1994) and glucose transporter-1 (Ebert et al., 1998). Oxygen dependent regulation of the HIF-1 alpha subunit appears to occur at multiple levels including mRNA expression, protein synthesis and stability and nuclear translocation (sem*nza, 1994).

Redox modification of protein sulphydryl groups has been demonstrated to directly modulate the DNA binding characteristics of a range of transcription factors including the fos/jun complex AP-1, NF-kappaB, the zinc finger protein Sp1 and the product of the c-myb oncogene (Wu et al., 1996).

ENERGY CRISIS

If retaliatory efforts to maintain normoxic levels of energy consumption do not suffice, which will happen if the oxygen supply is cut off completely as in anoxia and ischemia, then ATP manufacture does not satisfy ATP demand and high energy phosphate reserves are consumed. When ATP falls below about 50% of normoxic levels the membrane depolarizes and ion gradients are lost (Knickerbocker and Lutz, 2001). Since the loss of ATP is life threatening the emergency response is immediate.

The energy imbalance state is manifested by changes in the amount of energy metabolites. Initially phosphocreatine acts as an ATP buffer.

THE BRAIN

The brain is of special interest for hypoxia studies as it is extremely sensitive to reductions in oxygen supply. The reason for this vulnerability is the brain has committed high energy costs that cannot be compromised. For example, as much as 50–60% of the brain cells energy expenditure is devoted to transporting ions across the cell membranes in order to maintain cellular ion homeostasis (Lipton, 1999). As a result the brain suffers energy failure after only a few minutes interruption in oxygen supply.

Neural molecular adaptations to brain hypoxia/ischemia

The brain, however, has some ability to accommodate a limited exposure to hypoxia or ischemia, but there are clear neuronal subtype differences. For example, in mammalian brains the hippocampal CA1 pyramidal neurones have been found to be more sensitive to hypoxia than the CA3, CA4 or dentate gyrus neurones. Expression of hypoxia inducible genes, in particular c-fos, c-jun, bcl-x, bcl-2 and HSP-70, has been strongly correlated with this differential in survivability (Honkaniemi et al., 1996).

In the ischemic mammalian brain HSP-70 has been implicated in an endogenous neuroprotective mechanism (Li et al., 1993). Within 2 hr of the ischemic insult HSP mRNAs are induced in areas affected by ischemia. The temporal profile of HSP-70 protein expression follows the pattern of selective vulnerability in ischaemic brain, being seen first in CA1, then in CA3, cortex and thalamus and later in the resistant regions of the dentate granule cells after graded global ischemia (Simon et al., 1991).

Anoxic brain failure

The depletion ATP has important harmful consequences, including the failure of ATP dependant ion pumps, the activation of free radical formation by increasing xanthine levels, and maladaptive changes in phosphorylation states of different enzymes and structural proteins (Lipton, 1999). The failure of ion pumps allows a net movement of ions across the cell membrane down concentration gradients. The resulting anoxic depolarization leads to a chain of lethal consequences (for review see Lutz et al., 2002).

Following depolarization there is a massive loss of excitatory neurotransmitters such as glutamate, aspartate and dopamine into the extracellular space in amounts that have toxic effects (Lipton, 1999). There is also an excessive release of dopamine causing tissue injury (Globus et al., 1988).

The uncontrolled and explosive rise in intracellular Ca²⁺ is thought to be particularly harmful, causing multiple dysfunctional effects including the stimulation of phospholipid hydrolysis with a rise in harmful free fatty acids, particularly arachidonic acid that leads to the formation of free radicals (Lutz et al., 2002).

Anoxic brain survival

Although brain anoxia intolerance appears characteristic of vertebrates, there are a few species that survive prolonged brain anoxia. The brain of the freshwater turtle Trachemys scripta can survive at least 48 hr of anoxia in vivo during which time extracellular K⁺ rises only slightly, from 3 to 6 mM (Sick et al., 1982b). ATP levels are maintained for at least 6 hr of anoxia (Lutz et al., 2002) and electrical activity is depressed but not suppressed (Fernandes et al., 1997).

The most important compensation that the turtle brain makes to survive anoxia is lowering its energy consumption to such a degree (70–80%) that brain energy needs can be fully met by anaerobic glycolysis (Lutz et al., 1985). As a result, the turtle brain is able to maintain ATP levels and ionic gradients during anoxia and thus avoid the fatal consequences of energy failure.

Since like the mammal, ion pumping accounts for a considerable proportion (≈50%) of the normoxic turtle brain energy expenditure (Edwards et al., 1989) a reduction in membrane ion leakage can provide important energy savings (Lutz et al., 1985). Several studies indicate that channel arrest is indeed initiated in the anoxic turtle brain. During anoxia, potassium flux is significantly reduced (Chih et al., 1989; Pek and Lutz, 1998), there is a decrease in the density of voltage gated Na⁺ channels (Perez-Pinzon et al., 1992a) and a decrease in NMDA receptor activity (Bickler et al., 2000).

Important energy savings come from an almost full suppression in turtle brain electrical activity during anoxia, but this comes about in a complex manner (Fernandes et al., 1997). There is also a reduction in brain protein synthesis in the anoxic turtle (Frazer et al., 2001).

An intriguing question raised by these findings—does the turtle brain have altered molecular pathways of hypoxia sensing and response that account for its extraordinary ability to survive anoxia?

Hypoxia sensing

Neither the identity or sensitivity of any turtle hypoxia sensing mechanism is known, but interestingly, the in situ oxygen affinity of turtle brain cytochrome aa3 is much lower than that of the rat (Fig. 1). Lahiri has suggested that cytochrome aa3 may be an oxygen sensor, at least in the carotid body (Lahiri et al., 1995). Presumably a lower affinity cytochrome would signal hypoxia at higher PO₂ levels and may initiate hypoxia defense mechanisms well before the hypoxia “crisis” that the mammal cytochrome signals.

Modulation in levels of key molecular components

In initial studies by Southern blot analysis on turtle lymphocyte DNA we identified hom*ologues in the turtle genome for a range of coding sequences including those for HSP-70, c-fos, c-jun, bcl-2 and HIF-1 alpha (Nallaseth and Lutz, 1999; Lutz et al., 2002). In further studies using Northern blot analysis we have observed increases in levels of gene transcripts for c-fos and c-jun in hindbrain after 4 hr of anoxia with a continuing increase in c-jun transcripts over 24 hr. An elevation in c-jun mRNA was also observed in forebrain in the first 6 hr of anoxia. These observations are consistent with the findings of Greenway and Storey (2000) who demonstrated a 2 fold increase in c-fos protein in Western analysis of whole brain after 20 hr of anoxia. We observed no change in levels of transcripts encoding HSP-70 in turtle brain over 24 hr of anoxia by Northern blot analysis. By contrast employing in situ hybridization analysis and immunohistochemistry, Bickler and Buck (1998) found significant region specific elevations in levels of HSP-70 transcript and protein respectively after 11 hr of anoxia. HSP-70 mRNA levels were found to increase in cortex, midbrain and cerebellum. This data may point to an important role for HSP-70 in surviving anoxia.

Our analysis of HIF-1 alpha and HIF-1 beta RNA expression revealed increased levels of HIF-1 alpha and HIF-1 beta transcripts over 24 hr of anoxia in hindbrain with moderate increases in HIF-1 beta transcripts over 6 hr in forebrain.

Detection of hypoxia/anoxia dependent DNA binding activity of protein(s) was carried out using Electrophoretic Mobility Shift Assays (EMSA). DNA binding of NFkappaB was maximal at zero and six hours of anoxia in turtle brain (Fig. 2).

Physiological mechanisms

Ion channels

There is evidence that the anoxic turtle brain undergoes a reduction in ion membrane permeability (K⁺, Na⁺ and Ca²⁺) (Lutz et al., 2002; Bickler et al., 2000). While a reduction in ion permeabilities would produce energy savings the degree of channel arrest may be slight. A substantial decrease in ion conductance would be expected to produce a corresponding increase in membrane input resistance but Perez-Pinzon and coworkers (Perez-Pinzon et al., 1992b), found a slight decrease in input resistance in anoxic Purkinje cells of isolated turtle cerebellum, and Doll and colleagues (Doll et al., 1991) could detect no change in whole cell input resistance of turtle cortical pyramidal neurons during anoxia.

K⁺ channels

Chih et al. (1989) found that the rate of K⁺ efflux in the anoxic turtle brain was substantially lower than that seen in the normoxic brain, the first indication of “channel arrest” (Hochachka, 1986) in this system. There is evidence that this long term decrease in K⁺ leakage is partially mediated by adenosine receptors (Pek and Lutz, 1997).

K_ATP channels

There is wide speculation that ATP sensitive K⁺ channels which are activated during periods of energy failure, serve a protective function in the mammalian brain. In the turtle brain activated K_ATP channels are involved in the down-regulation of membrane ion permeability during the initial energy crisis period when ATP is depleted, but are switched off when the full anoxic state is established and tissue ATP levels have been restored (Pek and Lutz, 1998). K_ATP channels are also be critically involved in regulating dopamine release during anoxia (Milton and Lutz, submitted). However, in contrast to the mammal, K_ATP channels are not a major route for K⁺ efflux in the energy depleted turtle brain (Pek and Lutz, 1998).

Na⁺ channels

In support of the suggestion that hypoxia tolerant cells have inherent low channel densities (Hochachka, 1986), Edwards et al. (1989) found the density of voltage gated Na⁺ channel of turtle brains is 1/3 that of rat. By itself this difference cannot be an important factor in the 100 fold difference in anoxia tolerance. On the other hand changes in Na⁺ channel activity strongly affect metabolic rate. The only direct evidence for the modulation of voltage gated Na⁺ channels in the turtle comes from a study which found that 4 hr of anoxia produced 42% decline in the density of voltage gated Na⁺ channels in the isolated turtle cerebellum (Perez-Pinzon et al., 1992a). Interestingly, there is evidence that hypoxia reduces cell excitability in human neocortical brain slices by increasing the probability for Na⁺ channels to be in the inactive state (Cummings et al., 1993).

Ca²⁺ channels

The uncontrolled inflow of Ca²⁺ into mammalian neurons through, for example, the over-stimulated NMDA glutamate receptor, signals a wide variety of pathological processes and is thought to be one of the principal causes of anoxic brain death. Thus, a down regulation of Ca²⁺ channels could be of advantage during anoxia. The anoxia induced reduction in Ca²⁺ permeability appears to be adenosine mediated (Buck and Bickler, 1995). Bickler et al. (2000) have found that NMDARs are silenced by at least three different mechanisms operating at different times during anoxia. In pyramidal neurons from the turtle cerebrocortex, 1–8 min anoxia suppressed NMDAR activity by 50–60%. This rapid decrease in receptor activity was controlled by activation of phosphatase 1. A further decrease during 2 hr of anoxia was associated with an increase in intracellular Ca²⁺. A reversible removal NMDARs from the cell membrane was seen in turtles that had been held several days in anoxia at 3°C (Bickler et al., 2000).

Neurotransmitters and neuromodulators

Adenosine

In the turtle shortly after the onset of brain anoxia there is a substantial but temporary rise in extracellular adenosine (Nilsson and Lutz, 1992). Adenosine receptors mediate an increase in cerebral blood flow early in anoxia (Hylland et al., 1994), the reduction in membrane K⁺ leakage (Pek and Lutz, 1997) and NMDA down regulation (Buck and Bickler, 1995).

One reason why adenosine is so effective in the turtle brain is that unlike the mammal where a rapid reduction in adenosine A₁ binding sites occurs during even brief hypoxia/ischemia, no such deterioration is seen over 24 hr anoxia in the turtle brain (Lutz and Manuel, 1999).

Glutamate

In the mammalian brain anoxia produces a massive release of the excitatory neurotransmitter glutamate with toxic effects. By contrast, there is little or no increase in glutamate in the brain of anoxic turtles (Nilsson and Lutz, 1991; Milton et al., 2001).

Dopamine

An uncontrolled rise in extracellular dopamine has also been implicated as an important cause of pathogenesis in the hypoxic/ischemic brain (Globus et al., 1988). By contrast in the turtle extracellular dopamine stays at normoxic levels over 4 hr anoxia. Treatment with the specific dopamine transport blockers during anoxia resulted in a significant increase in dopamine over basal levels, indicating that dopamine was continuing to be released and taken up during anoxia (Milton and Lutz, 1998).

GABA

In contrast to excitatory glutamate there is a substantial release of the major inhibitory neurotransmitter GABA in the anoxic turtle brain (Lutz et al., 2002). It is suspected that the release of GABA has a protective function in anoxia (Lipton, 1999). There is an increase in GABA_A receptor number within 2 hr of anoxia which continues to increase for at least 24 hr (Lutz and Leone-Kabler, 1995). This up-regulation in GABA_A receptors may function to increase the effectiveness of the inhibitory action of GABA that is released during anoxia.

CONCLUSION

In conclusion we suggest that in animals hypoxia is signaled at three levels: an immediate systemic response which involves central and peripheral chemoreceptors, an immediate/chronic gene response initiated by cellular oxygen signals and an immediate emergency or crisis response signaled by changes in energy metabolite concentrations.

The fall in oxygen detected by central and peripheral (aortic and carotid body) chemoreceptors activate immediate systemic retaliatory responses to make up the deficit in oxygen supply. These include increased ventilation and increased lung oxygen extraction and increased cardiac output and tissue perfusion.

At the cell and organ level chronic hypoxia is detected by intracellular molecular oxygen sensors which signal through specific promoter elements the initiation of downstream adaptations for enhanced oxygen delivery such as increased tissue vascularization and increased red blood cell manufacture.

If these measures do not suffice the cell goes into an energy crisis as ATP stores are depleted. The crisis is detected and signaled by metabolic indicators such as a fall in ATP and an increase in extracellular adenosine. These act to initiate the processes for a severe down-regulation of ATP demand, through a series of coordinated metabolic shut downs. The future challenge is to link the molecular/genetic events with the physiological mechanisms.

1

From the symposium Plant/Animal Physiology presented at the Annual Meeting of the Society for Integrative and Comparative Biology, 3–7 January 2001, at Chicago, Illinois.

References

The Society for Integrative and Comparative Biology

Email alerts

Citing articles via

More from Oxford Academic