Hypoxia, i.e. decreased availability of oxygen occurs under many different circumstances and can be either continuous or intermittent. Continuous hypoxia such as that experienced during periods of high altitude leads to physiological adaptations, whereas chronic IH (intermittent hypoxia) associated with sleep-disordered breathing manifested as recurrent apneas leads to morbidity. The purpose of the present chapter is to highlight recent findings on cellular responses to IH. Studies on cell culture models of IH revealed that for a given duration and intensity, IH is more potent than continuous hypoxia in evoking transcriptional activation. IH activates HIF-1 (hypoxia-inducible factor-1), the immediate early gene c-fos, activator protein-1, nuclear factor κB and cAMP-response-element-binding protein. Physiological studies showed that HIF-1 plays an important role in chronic IH-induced autonomic abnormalities in mice. IH affects expression of proteins associated with neuronal survival and apoptosis, as well as post-translational modifications of proteins resulting in increased biological activity. Comparisons between continuous hypoxia and IH revealed notable differences in the kinetics of protein kinase activation, type of protein kinase being activated and the downstream targets of protein kinases. IH increases ROS (reactive oxygen species) generation both in cell culture and in intact animals, and ROS-mediated signalling mechanisms contribute to cellular and systemic responses to IH. Future studies utilizing genomic and proteomic approaches may provide important clues to the mechanisms by which IH leads to morbidity as opposed to continuous hypoxia-induced adaptations. Cellular mechanisms associated with IH (other than recurrent apneas) such as repetitive, brief ascents to altitude, however, remain to be studied.
Molecular oxygen is essential for the survival of mammalian cells because of its critical role in generating ATP via oxidative phosphorylation. Hypoxia, i.e. the decreased availability of oxygen, occurs under many different circumstances. Continuous hypoxia is experienced during periods at high altitude. Physiological systems adapt to continuous hypoxia, and the biochemical and molecular mechanisms underpinning the adaptations have been extensively investigated [1,2].
IH (intermittent hypoxia) is more often experienced than continuous hypoxia in people living at sea level. Chronic IH is experienced as a consequence of sleep-disordered breathing (manifested as recurrent sleep apneas) in 50% of premature infants , 5% of middle-aged men and 2% of women after menopause [4,5]. In this condition, transient repetitive apnea (cessation of breathing) results in periodic hypoxaemia [decreases in arterial blood PO2 (partial pressure of oxygen)]. Each episode of apnea lasts no more than tens of seconds and the frequency of apneas may exceed 60 episodes per hour with blood haemoglobin saturation of oxygen reduced to as low as 50% in severely affected patients. Recurrent apnea patients in addition to chronic IH also experience chronic intermittent hypercapnia [elevations in arterial blood PCO2 (partial pressure of carbon dioxide)]. A major advance in the field of apnea research is the demonstration that exposing experimental animals to chronic IH alone is sufficient to result in physiological changes similar to those described in recurrent apnea patients . Unlike continuous hypoxia, humans experiencing chronic IH as a consequence of recurrent apneas exhibit morbidity including development of hypertension, myocardial infarctions and stroke [4,5]. The purpose of this chapter is to summarize recent information on cellular responses to the paradigm that IH simulates recurrent apneas and compare these with the responses to continuous hypoxia.
Transcriptional activation during IH
It is being increasingly recognized that activation of specific genes is an important mechanism by which hypoxia triggers long-term adaptive responses . The following section describes the transcriptional changes associated with IH and highlights differences between the effects of continuous hypoxia and IH.
HIF-1 (hypoxia inducible factor-1)
The transcriptional activator HIF-1 is a global regulator of oxygen homoeostasis that controls multiple key developmental and physiological processes . Over 60 HIF-1 target genes have been identified, including those encoding EPO (erythropoietin) and VEGF (vascular endothelial growth factor) . HIF-1 is a heterodimeric protein that is composed of a constitutively expressed HIF-1β subunit and an oxygen-regulated HIF-1α subunit. HIF-1 activity is induced under conditions of continuous hypoxia as a result of a decreased rate of oxygen-dependent proline hydroxylation, ubiquitination and proteasomal degradation of the HIF-1α subunit . HIF-1 transcriptional activity is also regulated via oxygen-dependent asparginine hydroxylation that blocks co-activator recruitment .
HIF-1 activation by IH was studied in a cell culture model of IH wherein PC12 (pheochromocytoma-12) cells were exposed to alternating cycles of hypoxia (1.5% oxygen for 30 s) and re-oxygenation (20% oxygen for 4 min) . Medium PO2 decreased from 74±3 mmHg to 55±5 mmHg during each episode of hypoxia. Reintroduction of normoxic gas restored the medium PO2 to the pre-hypoxic value. Cell viability, pH and osmolarity of the medium were unaffected up to 120 cycles of IH . Under these experimental conditions, HIF-1α protein increased in a stimulus-dependent manner as the duration of IH increased from 10 to 30 to 60 cycles  To study HIF-1 transcriptional activity, PC12 cells were transfected with a reporter gene in which the expression of firefly luciferase was driven by a HIF-1-dependent HRE (hypoxia-response element) upstream of an SV40 promoter . IH resulted in significant stimulus-dependent HRE transcriptional activity. These observations suggest that IH, despite a modest decrease in medium PO2 (~20–25 mmHg) was effective in inducing HIF-1α protein expression and HIF-1 transcriptional activity, two necessary pre-requisites for HIF-1-dependent gene transcription.
On the other hand, when cells were exposed to 60 min of continuous hypoxia, which was equivalent to the hypoxic duration accumulated during 120 episodes of IH (30 s each episode), the medium PO2 decreased from 80±6 to 10±2 mmHg (a decrease of 70 mmHg) within 5 min after the onset of hypoxia and remained at this level during the entire period of hypoxic exposure. Despite the marked reduction in PO2, 60 min of continous hypoxia was ineffective in increasing either HIF-1α protein expression or activation of HRE . However, extending the hypoxic duration to 4 h did result in a significant increase in HIF-1α protein expression as well as HRE activation , as reported by others . These observations suggest that for a given duration and intensity, IH is more potent in activating HIF-1 than continuous hypoxia.
Mechanisms associated with IH-induced HIF-1α stabilization and activation
A recent brief communication by Yuan et al.  suggests that HIF-1α stabilization by IH involves increased protein synthesis and activation of rapamycin-sensitive mTOR (mammalian target of rapamycin) signalling. These investigators reported that rapamycin inhibits IH-induced HIF-1α stabilization and IH increases phosphorylated mTOR levels as well as downstream S6 kinases. Furthermore, the effects of IH on mTOR activation as well as on HIF-1α protein can be prevented by inhibitors of IP3 (inositol-3 phosphate) receptors and PLC-γ (phospholipase C-γ) as well as calcium chelators, indicating that mobilization of intracellular calcium is an important upstream signalling event required for IH-induced mTOR activation and subsequent HIF-1α accumulation. These observations are intriguing in that they are in striking contrast to the inhibition of the mTOR signalling pathway reported with continuous hypoxia . However, Hui et al.  reported that during continuous hypoxia, translation of HIF-1α can be stimulated by an influx of extracellular calcium and activation of the classical PKC (protein kinase C) and mTOR pathways in PC12 cells. Whether proline hydroxylation and subsequent ubiquination pathways, which are critical for HIF-1α stabilization during continuous hypoxia, also play a role during IH requires further study.
HIF-1 activation by continuous hypoxia requires activation of PI3K (phosphoinositide 3-kinase) . However, neither LY294002 nor wortmannin, inhibitors of PI3K, were able to block IH-induced HIF-1 transcriptional activity. On the other hand, the calcium chelator BAPTA-AM prevented IH-induced HRE activation, indicating the involvement of calcium signalling pathways in IH-induced HIF-1 transcriptional activity . CaMKs (calcium/calmodulin-dependent protein kinases) are downstream signalling molecules that participate in calcium-mediated gene regulation. A previous study reported that continuous hypoxia leads to a transient (15 min) and modest (1.5-fold) increase in CaMKII activity in PC12 cells . In striking contrast, IH resulted in an exponential and nearly 6-fold increase in CaMKII activity in response to increasing cycles of IH and was associated with increased phosphorylation of CaMKII protein . Co-transfection of constitutively active CaMKII-290 mimicked the effects of IH. Inhibitors of either calmodulin (W-7) or CaMK (KN-93) prevented IH-induced but not continuous hypoxia-induced HIF-1 transcriptional activity . A CaMK inhibitor, however, was ineffective in preventing IH-induced HIF-1α protein expression, suggesting that CaMK-dependent signalling is critical for IH-induced HIF-1 transcriptional activation but not for HIF-1α protein expression. IH stimulated the CAD (C-terminal transactivation domain) of HIF-1α via a mechanism that is independent of Asn830 hydroxylation and CaMKII activation was required for this response . On the other hand, Asn830 hydroxylation is required for continuous hypoxia-evoked CAD activation . How might CaMKII contribute to IH-induced HIF-1 transactivation? Several lines of evidence suggest that p300/CBP proteins are the major co-activators for IH-induced HIF-1 activation [19,20]. IH increased p300 transcriptional activity, which was blocked by KN-93 . The signalling pathways associated with HIF-1 activation by IH are schematically presented in Figure 1. These observations suggest that IH-induced HIF-1 transactivation requires a novel signalling pathway involving CaMKII-dependent activation of p300 co-activator, suggesting striking differences in the signalling pathways associated with HIF-1 activation in response to IH compared with continuous hypoxia.
Functional significance of HIF-1 activation by IH
Patients with IH caused by recurrent apneas develop autonomic abnormalities including systemic hypertension, heightened ventilatory response to hypoxia and persistent activation of the sympathetic nervous system . IH-induced autonomic abnormalities have been attributed to activation of the carotid body, the primary sensory organ for detecting systemic hypoxia . Previous studies have shown that rodents exposed to IH also develop systemic hypertension, elevated plasma catecholamines, augmented hypoxic ventilatory response and elevated sympathetic nerve activity  similar to that reported in recurrent apnea patients. Supporting a role for the carotid body, it has been shown that chronic IH sensitizes chemoreceptor response to acute hypoxia and induces long-lasting sensory excitation [sensory LTF (long-term facilitation)] in rats . Alterations in the carotid body activity may contribute to increased sympathetic tone and systemic hypertension.
Complete HIF-1α deficiency results in embryonic lethality at mid-gestation, whereas hif-1α+/– HET (heterozygous) mice, which are partially deficient in HIF-1α expression, develop normally and are indistinguishable from WT (wild-type) littermates under normoxic conditions . Using hif-1α+/– HET mice, a recent study examined the importance of HIF-1 activation in IH-induced physiological alterations . In contrast to WT mice, IH was ineffective in evoking HIF-1α protein expression in HET mice . More importantly, the absence of a HIF-1α response to IH in HET mice was associated with a striking absence of blood pressure elevation, plasma catecholamines and augmented ventilatory response to hypoxia as well as an absence of carotid body sensitization and sensory LTF .
Clinical studies indicate that recurrent apneas are associated with increased serum levels of EPO and VEGF [26,27], two target genes activated by HIF-1. The effects of IH-induced elevations in circulating EPO on cardiac responses to ischaemia–reperfusion have been examined [28,29]. Exogenous EPO prevented deleterious effects of ischaemia–reperfusion in the heart . Studies in WT mice have shown that EPO expression was elevated by IH with a concomitant cardiac protection against ischaemia–reperfusion injury . In contrast, such cardiac protection against ischaemia–reperfusion was absent in IH-exposed hif-1α+/– mice, demonstrating a critical involvement of HIF-1 in these responses . These observations, taken together with the biochemical data, suggest that IH activates HIF-1-mediated transcription which appears to be critical for evoking cardio-respiratory responses. However, future studies are needed to identify the HIF-1-regulated downstream genes other than EPO and VEGF that contribute to IH-evoked cardio-respiratory changes.
Immediate early genes
Hypoxia is a potent activator of the c-fos gene, a member of the immediate early gene family . c-Fos protein in combination with c-Jun, another member of the immediate early gene family, forms the heterodimeric AP-1 (activator protein-1) complex. AP-1 functions as a transcriptional activator that drives expression of a variety of genes. Greenberg et al.  reported up-regulation of c-Fos protein in the central nervous system of rats exposed to 30 days of IH. Yuan et al.  examined the effects of IH on c-fos mRNA expression in PC12 cells. They reported that IH increases c-fos mRNA expression in a stimulus-dependent manner. The IH-induced increase in c-fos mRNA was in part due to increased c-fos transcriptional activation as evidenced by a reporter gene assay. Point mutations in the c-fos promoter showed that SRE (serum responsive element), CRE (calcium response element) but not FAP [AP-1/CRE (cAMP-response-element)-like cis-element] are critical for IH-induced c-fos promoter activation . Interestingly, the magnitude of c-fos activation by IH was dependent on the duration of re-oxygenation between the hypoxic episodes, and not on the duration of the hypoxic episodes . Exposure to continous hypoxia (comparable cumulative duration of hypoxia accumulated during IH) had virtually no effect on c-fos expression . These observations reiterate that for a given duration, IH is a more potent transcriptional activator than continuous hypoxia. Interestingly, following termination of IH c-fos mRNA remained elevated for at least 3 h, whereas it returned to control levels within 1 h of terminating continuous hypoxia , demonstrating another striking difference between both forms of hypoxia. The long-lasting c-fos mRNA activation by IH is reminiscent of IH-induced long-lasting activation of the carotid body sensory activity  and breathing [32,33] reported in intact animals. These findings indicate that long-lasting activation, which is a hallmark of IH in intact animals, can be elicited by IH even at the gene level. The mechanisms by which IH leads to long-lasting gene activation, however, remain to be studied.
What might be the functional significance of IH-evoked c-fos activation? IH resulted in increased AP-1 transcriptional activity and expression of mRNA of TH (tyrosine hydroxylase) a downstream AP-1-regulated gene . Co-transfection with antisense c-fos abolished IH-induced AP-1 activation and TH mRNA expression, implying that c-fos contributes to IH-induced AP-1 activation and downstream gene transcription . Since TH is the rate-limiting enzyme in catecholamine synthesis, it is likely that IH-induced TH activation contributes in part to the increased catecholamine levels reported in IH-exposed animals , as well as in humans experiencing IH due to apneas .
NF-κB (nuclear factor κB)
NF-κB is an important transcriptional regulator of inflammatory mediators. Ryan et al.  reported that in obstructive sleep apnea patients NF-κB activity in monocytes is increased resulting in elevated serum TNFα (tumour necrosis factor-α) and IL-8 (interleukin-8) levels. Elevated TNFα levels correlated with arterial oxygen desaturation, and treating apnea patients with CPAP (continuous positive airway pressure) normalized TNFα levels . These observations led to the suggestion that IH activates NF-κB. Consistent with this possibility is the demonstration that IH stimulates NF-κB-mediated transcriptional activation in HeLa cells . In addition, Greenberg et al.  found elevated NF-κB activity in IH-exposed mice and apnea patients. The mechanisms by which IH activates NF-κB, however, remain to be investigated.
The studies described thus far suggest that IH is a potent activator of HIF-1, c-fos and NF-κB. Whether IH activates other transcriptional factors requires further study.
Altered protein expression
The following section summarizes the effects of IH on protein expression and post-translational modifications.
Protein expression leading to neuronal injury and apoptosis
The influence of IH on acid-base transporters in the mouse central nervous system was investigated by Douglas et al. . In this study, 2- to 3-day-old mice were exposed to alternating cycles of 2 min of hypoxia (6.0–7.5% oxygen) and 3 min of normoxia (21% oxygen) for 8 h/day for 28 days. IH exposure decreased the expression of sodium/hydrogen exchangers (the NHE-1 and 3 isoforms) and sodium-bicarbonate co-transporter proteins in the central nervous system, especially in the cerebellum. It has been proposed that IH, by down-regulating the acid-extruding capacity of neurons, renders them more susceptible to acidic insult and subsequent neuronal injury.
Recent studies suggest that IH leads to cognitive deficits and is associated with neuronal apoptosis in experimental animals . Acute exposure to IH (10% oxygen followed by 21% oxygen for 90 s each) for 6 h resulted in a much greater injury of the CA1 region of the hippocampus than the CA3 region. Proteomic analysis of the CA1 and CA3 regions of the hippocampus derived from IH and control rats demonstrated a complex pattern of protein alterations that accompanied CA1 injury . IH augmented the expression of stress-induced proteins, primarily chaperone proteins and proteins related to apoptosis. In addition, comparison of protein maps of the CA1 and CA3 regions revealed a marked up-regulation of a hitherto uncharacterized protein in the CA1 region but not in the CA3 region, which might play a key role in CA1 injury during IH .
Exposure of rats to alternating cycles of hypoxia (10% oxygen for 90 s) and normoxia (21% oxygen for 90 s) for up to 14 days has been shown to selectively up-regulate protein expression and activity of COX (cyclo-oxygenase)-2, but not COX-1, in cortical regions of the brain . The up-regulation of COX-2 was associated with increased neuronal apoptosis and neurobehavioural abnormalities as evidenced by deficits in the acquisition and retention of a spatial task . Furthermore, studies by Goldbart et al.  suggest that decreased phosphorylation of CREB (cAMP-response-element-binding protein) accounts for decreased neuronal survival and spatial learning defects observed in IH exposed rats. Gozal et al.  showed that in PC12 cells, IH induced an earlier and more extensive apoptotic response than continuous hypoxia that was associated with enhanced caspase activity.
Post-translational modifications of proteins
Hypoxia, in addition to affecting protein expression, also alters the function of existing proteins via post-translational modifications of specific amino acid residues involving phosphorylation and dephosphorylation reactions. Studies on PC12 cells showed that IH activates a variety of protein kinases including MAPKs [mitogen-activated protein kinases; ERK1/2 (extracellular-signal-regulated kinase 1/2) and Jun kinase], PKC and CaMKII . Comparison of continuous hypoxia with IH revealed some notable differences in protein kinase activation. For instance, IH resulted in a sustained increase in CaMKII (5.5-fold) , whereas continuous hypoxia caused a transient (15 min) and modest increase (1.5-fold activation) . Likewise, Jun kinase was only activated by IH but not by continuous hypoxia . Continuous hypoxia but not IH activates p38 kinase in PC12 cells [12,18]. Although both IH and continuous hypoxia activate ERK-1/2 in PC12 cells, ERK-mediated signalling is critical for HIF-1 activation by continuous, but not by IH . Thus, although both IH and continuous hypoxia activates similar protein kinases, the kinetics of activation and the downstream targets seem to differ between the two forms of hypoxia.
In PC12 cells IH increased the activity of the TH enzyme, which is in part due to increased phosphorylation of the Ser40 residue . IH-induced TH activation could be prevented by inhibitors of CaMK and PKA (protein kinase A) . Gozal et al.  compared the changes in TH mRNA, protein expression and activity in various brain regions of male rats exposed to continuous hypoxia and IH, and concluded that differences in TH activity between IH and continuous hypoxia are related to different phosphorylation patterns. Studies by Goldbart et al.  showed that IH results in a time-dependent decrease in phosphorylation and nuclear binding of CREB without changes in total CREB protein. It will be of interest to determine whether other types of post-translational events like sulfation, glycosylation and ubiquitination also play a role in altered protein activities associated with IH.
ROS (reactive oxygen species) signalling during IH
It is evident from studies outlined above that IH and continuous hypoxia have profoundly different effects on transcriptional activation, signalling pathways and protein expression. What makes IH different from continuous hypoxia? A major difference is that IH is interspersed with periods of re-oxygenation, which are absent during continuous hypoxia. In cell cultures the duration of re-oxygenation rather than the hypoxic phase was found to influence IH-induced gene expression . Thus it was proposed that IH might generate ROS during the re-oxygenation phase analogous to that reported with ischaemia–reperfusion.
ROS generation during IH
Aconitase is an enzyme in the tricarboxylic acid cycle and is present both in the cytosol and in mitochondria. ROS inhibits aconitase activity and antioxidants prevent this response. Thus inhibition of aconitase activity can serve as an index of ROS generation . Aconitase activity was markedly down regulated in the carotid body  and in adrenal medulla  from IH-exposed rats as well as in IH exposed PC12 cells . Likewise, TBARS (thiobarbituric acid reactive substances), an indicator of protein oxidation are elevated in tissues from IH-exposed animals and antioxidants prevented this response . Taken together, these observations suggest that IH increases ROS generation. However, these studies do not indicate which of the reactive oxygen species (i.e. O2−, H2O2 or OH−) are elevated in response to IH.
ROS generation during IH could involve inhibition of the mitochondrial electron transport chain complex as well as activation of several oxidases. IH decreases mitochondrial complex I but not complex III activity in cell cultures . On the other hand, there is evidence suggesting that continuous hypoxia generates ROS by inhibiting complex III of the electron transport chain . The mechanisms by which IH selectively targets complex I and the potential contribution of oxidases to IH-induced ROS generation remain to be investigated. MnTMPyP [manganese (III) tetrakis-(1-methyl-4-pyridyl)-porphyrin pentachloride], a superoxide dismutase mimetic that scavenges superoxide anion without generating H2O2, prevented up regulation of c-fos and AP-1 activation in PC12 cell cultures exposed to IH . Likewise, MnTMPyP prevented IH-induced HIF-1α accumulation . Furthermore, the observation that MnTMPyP also prevents IH-induced activation of Jun kinase  suggests a role for ROS in IH-induced protein kinase activation. ROS might activate protein kinases either directly or indirectly by inhibiting protein phosphatases or a combination of both, clarification of which requires further study. Physiological studies have shown that MnTMPyP as well as N-acetyl cysteine, a precursor of glutathione (another antioxidant), prevent IH-induced elevations in blood pressure, plasma catecholamines  and augmented ventilatory response to hypoxia as well as functional alterations in carotid body activity . These studies suggest that ROS signalling plays an important role in mediating biochemical and physiological responses to IH (Figure 2).
In summary, the studies described above suggest that IH, such as that associated with recurrent apneas, affects a variety of biochemical processes, including transcriptional activation of genes and signalling mechanisms involving post-translational modification of proteins that may underlie the physiological responses. Emerging evidence indicates that ROS-mediated signalling is critical for evoking IH-induced cellular and systemic responses. However, it must be pointed out that our understanding of the physiological and cellular responses to IH is in its very early stages. There are several gaps in our knowledge as to the mechanism by which IH affects gene and protein expression. Except for a few proteins, there is a paucity of information regarding whether hypoxia-induced up-regulation of proteins is due to de novo synthesis or to increased protein stabilization, or to translocation of proteins to specific cellular compartments. Thus far studies examining the influence of IH have focused on those proteins that show increased expression. But it is equally important to assess the role of proteins whose expression is down-regulated during IH. Future studies using genomic and proteomic approaches may provide important clues to the process by which IH leads to morbidity as opposed to the adaptations induced by continuous hypoxia. In addition to recurrent apneas, humans also experience IH during repetitive brief ascents to high altitude, wherein each hypoxic episode lasts for a few hours . It remains to be determined whether both forms of IH activate similar cellular mechanisms.
• Hypoxia (i.e. a decrease in arterial blood oxygen) can be continuous or intermittent.
• Continuous hypoxia such as that experienced during sojourns at high altitude leads to physiological adaptations, whereas chronic IH, associated with recurrent apneas, results in morbidity.
• For a given duration and intensity, IH is more potent than continuous hypoxia in activating various transcription factors including HIF-1 and c-fos.
• Physiological studies revealed that HIF-1-mediated transcription plays a critical role in mediating IH-induced cardio-respiratory responses.
• The kinetics of protein kinase activation and the type of kinase activated by IH differ considerably from continuous hypoxia.
• IH increases ROS generation and ROS-mediated signalling mechanisms play a critical role in IH-induced cellular and systemic responses.
We are grateful for the grant support from the National Institutes of Health, Heart, Lung and Blood Institute (HL-25830 and HL-46462).
- © The Authors Journal compilation © 2007 Biochemical Society