INTRODUCTION

TOP ABSTRACT INTRODUCTION THE CELL CYCLE THE DIRECT EFFECT OF... SUMMARY REFERENCES

ENVIRONMENTAL STRESSES CAN have direct effects on cell cycle activity. Ultraviolet (UV) and ionizing (IR) radiation can cause DNA damage that leads to dysplasia or even morbidity and mortality of the embryo (138). Other factors such as heat, overcrowding (contact inhibition), and dessication can lead to cell cycle arrest, resulting in either temporary suspended animation or eventual cell death. Oxygen (O₂) deprivation or hypoxia is another such stress that appears to have a direct effect on cell cycle activity.

O₂ deprivation is not only important in pathogenetic mechanisms of disease processes but is also involved in development and cell growth in early life, especially during embryogenesis (59, 159, 160). To illustrate the former, consider some of the disease conditions that we regularly encounter or some of the natural physiological stresses that are often imposed on organisms. A critical component of respiratory and cardiovascular diseases in children and adults as well as life at high altitude is the exposure to periodic or continuous episodes of hypoxia. These episodes, whether acute or chronic, can result in pulmonary hypertension and neurological and cardiovascular disorders (72, 95, 178, 179). Prolonged ischemia or hypoxia in fetuses or in neonates can cause irreversible brain damage and injury to O₂-sensitive organs such as the lung, heart, and kidney (22, 79, 132, 166). Although such conditions are not clinically infrequent, the cellular and molecular events that take place in and outside cells are not well understood, mainly due to a lack of a suitable animal model.

Mammals are normally extremely intolerant of severe or prolonged hypoxic exposure, and O₂ deprivation in utero can often ensue in fetal demise. It has also been demonstrated clinically that patients who live for prolonged periods of time with desaturated hemoglobin (such as with congenital heart disease) show stunted growth and development in early life (73) that is not likely related to dietary intake. Similar growth-related problems have also been discovered for high-altitude sojourners (87). Therefore, low-O₂ conditions may not only lead to injury but may also affect growth rate early in life.

However, several species, such as the turtle and brine shrimp, demonstrate imperviousness to lack of O₂ (26, 88). We and others have previously shown that the fruit fly, Drosophila melanogaster (D. melanogaster), embryo or adult is also capable of surviving and recovering from several hours of prolonged hypoxia (42, 74, 176). The understanding of the cellular and molecular mechanisms underlying the response to hypoxia in this and other species has been essentially limited. It has been reported that hypoxia can lead to either hypoplasia as seen in the growth retardation reported in response to in utero hypoxia or hyperplasia as demonstrated in the pulmonary vasculature. How such a stress can produce contrasting effects is still under intensive experimentation, and the molecular mechanisms that permit these clearly varied responses to lack of O₂ are not completely understood but are also being aggressively investigated. This review will address the advances that have been made in our understanding of the cellular hypoxia response in the recent past.

	THE CELL CYCLE

TOP ABSTRACT INTRODUCTION THE CELL CYCLE THE DIRECT EFFECT OF... SUMMARY REFERENCES

An understanding of the regulation of the cell cycle is requisite in the attempt to determine how hypoxia might affect cell cycle activity. The mechanism of cellular duplication and replication is regulated by a complex hierarchy of genetic and metabolic networks. In the early 1970s, a rash of activity led to the elucidation of the essential underpinnings of the cell cycle machine, i.e., the oscillation of certain cell cycle proteins, namely the cyclins and their partners, the cyclin-dependent kinases (cdks), in a phasic, cyclical fashion (55). Cell cycle periodicity can also be viewed as linear sequential events that are iterated as needed by the organism (137). Rapid iterations of cell cycle activity occur in the early embryo and may be as short as 10 min per cycle. On the other hand, cell cycle length within individual tissues can be as long as 48 h per cycle, depending on the species and the developmental stage. Finally, a critical component of the cell cycle is the ability to cease replicative activity and to become quiescent. Cells can then reenter the cell cycle if the proper stimulus is provided or replication becomes deranged, as in cancer. Control of the cell cycle therefore involves modulation of several transition states and stages of varied length. The traditionally defined stages of the cell cycle are 1) S phase, during which DNA replication occurs; 2) gap phase 2 (G₂), during which proteins required for mitosis are accumulated; 3) M phase (mitosis or maturation stage), during which chromatin condensation, nuclear envelope breakdown (NEBD), chromatid separation, and cytokinesis occur; 4) gap phase 1 (G₁), during which genes necessary for DNA replication are activated and the protein agents of S phase progression are accumulated; and 5) G₀, during which cells can exit the cell cycle and enter a state of differentiation or quiescence (Fig. 1).

View larger version (48K): [in this window] [in a new window]   Fig. 1.   Analog diagram representing a simplified version of the cell cycle, with a focus on the G1/S phase transition. A minimum number of regulatory cell cycle proteins are depicted in this figure. Cyclin D/cyclin-dependent kinase (cdk)4,6 is activated to induce initiation of S phase before cyclin E; however, both complexes are required. Cyclin D, in cooperation with cyclin E/cdk2, hyperphosphorylates retinoblastoma protein family (RBF), causing the release of RBF from the dimer elongation factor (E2F)/dimerization partner (Dp), a transcription factor that activates several S-phase-requiring genes. Cyclin A/cdk2 also participates in progress through S phase. In response to DNA damage, several regulatory pathways are activated to prevent entry into S phase. Ataxia telangiectasia mutated (ATM) and Chk1 or Chk2 are variably capable of inducing a G1/S phase block by inhibiting cdk2, cyclin E, or E2f/Dp. The embryonic fruit fly hypoxia-responsive segment of the cycle (interphase arrest) is delineated by the central arrowhead. G0, quiescent or differentiated state; G1, gap phase 1; S phase, DNA synthesis; G2, gap phase 2. Components of mitosis are as follows: P, prophase; PM, prometaphase, M, metaphase; A, anaphase; T, telophase; CK, cytokinesis.

Control of cell cycle activity can be exerted both during the various stages of the cell cycle and at the points of transition from state to state. Thus cell cycle activity can also be modulated at the G₁/S and G₂/M transition points. Environmental cues can also impact on cell cycle activity so that there are both endogenous and exogenous cell cycle regulatory factors that can modulate both the rate of cell cycle activity and the stage in which the cell must reside. External/extracellular modulation of cell cycle activity involves both mitogenic and pathological signals. Therefore, there are also intrinsic and extrinsic conditions that can result in reentry into the cell cycle or exit from the cycle, i.e., arrest or quiescence. Some of these extrinsic factors include contact inhibition, nutrient supply, cell size, temperature, and a variety of stresses such as hypoxia. We are particularly interested in the effect of environmental hypoxia on cell cycle activity and cell injury.

History

In vitro cells have provided the most information about cell growth and cell division. Initially, studies of cell cycle activity were performed in rather large cells of the unicellular protozoan, Physarum polycephalum (96), in which cells at different stages of the cell cycle were fused. These studies demonstrated that late G₂- or M-phase cells contained an M-phase promoting factor (MPF) that could accelerate the entry into mitosis in early G₂ cells. Subsequently, it was reported that the fusion of HeLa cells at different stages of the cell cycle induced G₂-phase nuclei to delay entry into mitosis until the S-phase nuclei had completed DNA synthesis and that G₂-phase nuclei could not initiate DNA replication when fused to S-phase cells (153). This indicated that mitotic- and S-phase-inducing substances or an S-phase promoting factor (SPF) was contained in the cytosol of these cells at particular stages of the cell cycle. Additionally, researchers reported earlier that certain substances in the cell increased to a critical level to induce mitosis (55). The oscillations in the cell cycle were first noted in Xenopus laevis eggs (100), and the determination that an activity in frog eggs could induce M phase in immature G₂ oocytes (124) led to identification of MPF. The activity of MPF was demonstrated in frog oocytes in which it was noted that MPF was able to induce meiotic metaphase entry in G₂-arrested cells; in addition, mitotic entry occurring in the absence of protein synthesis was also observed.

A major step in the understanding of cell cycle progression and regulation was the purification of the MPF or mitosis promoting factor (also known as the M phase-specific promoter and meiosis promoting factor). The kinase activity of MPF was shown to oscillate in phase with the major transition points of the cell cycle and be absolutely required for transit through the cycle (118). Even more significant progress in understanding cell cycle progression was made when the composition of MPF was determined to be the complexing of a protein kinase cdk1, also known as histone 1 kinase or cell division cycle (cdc) 2, to an activating, regulatory subunit, cyclin B. Another major breakthrough occurred when it was determined that the catalytic or enzymatic element (cdk1) of MPF was the cdc2 gene of Schizosaccharomyces (S.) pombe and cdc28 gene in Saccharomyces (S.) cerevisiae and is required for the G₁/S transition (for reviews, see Refs. 60 and 66). Human cdk1 was later discovered and found to possess a G₂/M role. There is apparently a single cdk (cdc2, cdc28) for the major transitions in yeast cells, whereas the G₁/S transition in higher eukaryotes is controlled by different cdk homologs (58). In mammals, different cyclins or cdks assemble at different points in the cell cycle (reviewed in Refs. 92, 129, 161).

Originally, it was assumed that cyclin A was also in a complex with cdk1 (cyclin A/cdk1) and that this complex appeared to oscillate with the same cell cycle kinetics as MPF (cyclin B/cdk1) (46). It was also noted that both cyclin A and cyclin B were degraded at the completion of metaphase. In addition, cyclin B degradation is required for the exit from mitosis. It was therefore determined that the protein kinase cdk1 operated at essentially two major transition points in the cell cycle, G₁/S and G₂/M, and that phosphorylation and dephosphorylation events were integral components of cell cycle progression. A simple model of the cell cycle was thus generated in which the complexing of cyclin A to cdk1 initiated S phase entry and cyclin B and cdk1 was required for the initiation of mitosis (Figs. 1 and 2).

View larger version (44K): [in this window] [in a new window]   Fig. 2.   Analog diagram representing a simplified version of the cell cycle, with a focus on the G2/M transition. A minimum number of regulatory cell cycle proteins are depicted in this figure. Both cyclin A/cdk2 and cyclin B/cdk1 are required for progress through G2 and entry into mitosis. The activating cdc25 phosphatase and the inhibitory wee1 kinase essentially control the activation of cdk1 and possibly cdk2 by phosphorylation and dephosphorylation events. Inactivation of cdks and destruction of mitotic cyclins, first cyclin A and then cyclin B, are required for the metaphase-to-anaphase transition as well as exit from mitosis. Activation of the anaphase-promoting complex (APC/C) is also required for the metaphase-to-anaphase transition and mitotic exit, whereby the APC/C serves to sequentially degrade cyclin A/cdk2, sister chromatid cohesive factors such as the securins, and cyclin B/cdk1 under the control of cdc20 at the metaphase-to-anaphase transition and by Hct1 beyond anaphase and during early G1. DNA damage or spindle abnormalities activate the spindle checkpoint, which functions via inhibition of the APC/C by several elements inclusive of the Mad and Bub proteins. The embryonic fruit fly hypoxia-responsive segment of the cycle (metaphase arrest) is depicted by the large central arrowhead.

Unraveling the mysteries associated with progression through the cell cycle was hampered by the apparently promiscuous nature of these proteins (e.g., cdk1 and its multiple partners, cyclins A and B), making it difficult to specify the roles that each played in the cell cycle. However, three characteristics of the cell cycle eventually led to the clarification of the roles of these proteins in cell cycle regulation: 1) sequential cell cycle events required the completion of a prior step to be initiated, 2) new protein synthesis was required to overcome meiotic arrest points in oocytes and to initiate any new mitotic cycle, and 3) protein degradation was an essential component of progress through the cycle (100).

Regulation of The Cell Cycle: Modern View

The number of proteins that are known to participate in the regulation and execution of the cell cycle has grown exponentially during the past two decades. Below, we will present a simplified overview of what is currently known about cell cycle regulatory proteins and the stages of the cell cycle that they impinge on. It should be noted that many outstanding, in-depth reviews have been written that report on each stage of the cell cycle.

Cells in G₀ can be stimulated to return to the cell cycle and thereby reenter G₁ (146). Factors such as mitogens can trigger cyclins D and E to initiate the transcription of genes and the accumulation of protein products necessary to initiate DNA synthesis and chromosome duplication and traverse the G₁/S boundary. Clearly, there are several stages of the cell cycle involved, and below we describe them as well as some of the key elements that regulate them.

G₁/S transition. During G₁, several transcriptional and translational events are coordinated so as to initiate DNA replication. Cells are normally induced to leave G₁ and enter S phase by mitogenic agents such as growth factors. These extracellular signals are transduced by "membrane receptors" that activate the translation of the D-type cyclins (cyclins D1, D2, D3 in mammals) and their cognate protein kinase partners, cdk4 and cdk6 as well as cyclin E and cdk2 (Fig. 1).

S phase. The retinoblastoma protein product (RBF or pRb) is the most well studied of the cell cycle regulatory proteins. It was first noted to be absent or inactivated in eye tumors (retinoblastomas) and is also found in small cell lung cancer, sarcomas, and bladder cancer (91). RBF is targeted by mammalian viral tumor proteins such as E1A from adenovirus, T antigen from simian virus 40, and E7 from human papilloma virus, which binds directly to inactive, hyperphosphorylated RBF (49, 89, 128) and therefore enhances E2F activity leading to an accelerated G₁/S. The hypophosphorylated form of RBF is the active form found bound to E2F1/Dp in quiescent cells and in early G₁ and thereby exerts its repressive function. During mid- to late-phase G₁, RBF is hyperphosphorylated and removed from the E2F1/Dp dimer, therefore inactive and permitting the initiation of S phase by E2F1/Dp. Later in the cycle, Cyclin D targets RBF to be dephosphorylated by the phosphoprotein phosphatase type 1 so that it can once again repress E2F1/Dp activity through complexation (121). Therefore, cells without RBF or with compromised RBF function do not have cyclin D regulation and demonstrate increased oncogenic potential.

Progress through S phase. Cyclin A, whose kinase activity was detailed in S. pombe, is synthesized in late G₁. Although cyclin A is synthesized late in G₁, its kinase activity can be detected in the S phase. The microinjection of anti-cyclin A antibodies or anti-sense oligonucleotides directed against cyclin A inhibits DNA synthesis, indicating that cyclin A is required for DNA synthesis (39, 108). The cyclin A/cdk2 complex localizes to nuclear/chromosomal replication foci (16). However, cyclin A also demonstrates peak kinase activity in G₂ and is required for entry into mitosis. Cyclin A binds cdk2, whose peak kinase activity occurs in G₂. Cyclin A is suddenly degraded thereafter. Therefore, cyclin A/cdk2 activity is required both at the G₁/S and G₂/M transition points (143).

G₂/M transition. Originally, it was thought that entry into mitosis is triggered solely by MPF, which is composed of cdk1 and cyclin B. However, soon it was appreciated that entry into, progress through, and exit from mitosis are catalyzed by the accumulation and activation of both cyclin B/cdk1 and cyclin A/cdk2. Cyclin B/cdk1 is absolutely required for mitotic entry, whereas cyclin A/cdk2 appears to be necessary for the entry into prophase as well as progress through mitosis. Under the dually opposing control of cdc25, a phosphatase that activates cdk1, and wee1, a kinase that inhibits cdk1, initiation of mitosis is achieved by accumulating cyclins A and B while inhibiting wee1 kinase and disinhibiting cdc25 (15) (Fig. 2).

Mitosis. PROPHASE. Furuno et al. (61) suggested that cyclin A/cdk2 is able to move cells through most of prophase but that cyclin B/cdk1 activity is required at the end of prophase for NEBD. Cyclin A/cdk2 is therefore necessary for the initiation of chromosome condensation, disassembly of nucleoli, and the activation and perhaps translocation of cyclin B/cdk1 into the nucleus. On activation and translocation of cyclin B/cdk1 to the nucleus, cells enter metaphase, which involves the completion of chromosome condensation, NEBD, and spindle formation.

Cdk inhibitors. Another level of cell cycle regulatory control involves inhibitors of cdk activity, the cdk inhibitors (CDKIs). These include members of the CIP/KIP family (p21, p27, p57), which are presumed to inhibit all G₁-phase cyclin/cdk complexes, and members of the INK4 family (p16, p15, p18, p19), which are thought to inhibit only cyclin D/cdk complexes (reviewed in Ref. 183). On activation by p53 during a variety of stresses, p21 (Cip1, Waf1, or Sdi1) is able to bind to and directly inhibit a number of cyclin/cdk complexes including cyclin E/cdk2 and cyclin A/cdk2 but not cyclin B/cdk1 (Figs. 1 and 2) (80). p21 also prevents S phase and G₂ cells from entering mitosis and causes early prophase cells to return to interphase (G₂) (61). Mutations in the CDKIs have been implicated in several tumors (98).

Checkpoints

The DNA replication and damage checkpoints. Cell division is controlled by checkpoint controls that ensure that DNA is completely replicated (DNA replication checkpoint) and undamaged (DNA damage checkpoint) before cell division occurs (9, 54). The DNA damage checkpoint activates pathways that delay cell cycle progression to allow for the repair of damaged DNA (82). The G₁ DNA damage checkpoint arrests cells in G₁ before DNA synthesis, whereas the G₂ DNA damage checkpoint arrests cells in G₂ before mitosis (Figs. 1 and 2) (185). Activation of the DNA damage checkpoint involves the phosphoinositide kinase homologs ataxia telangiectasia mutated (ATM) and ataxia telangiectasia-related protein (ATR), which functions at both G₁ and G₂. The DNA damage checkpoint imposes a block at the G₁/S transition point by inducing ATM, ATR, chk1 and 2 protein kinases, and p53 (Fig. 1), which subsequently activates p21 (51). Integral to the G₁ arrest is the p53 transcription factor (13, 99) (Figs. 1 and 3). During exposure to IR and UV, p53 is stabilized by ATM and activates many genes and the CDKI p21 (12, 52, 173), which blocks transcription of cdk complexes required for S-phase entry. There are, however, p53-independent mechanisms that are also required for G₁ arrest in response to DNA damage (1). It is interesting to note that recent studies have linked p53 and hypoxia-inducible factor 1 (HIF1) activity, indicating that p53 is able to directly bind to HIF1 and destabilize its expression by promoting ubiquitination mediated by mouse double minute chromosome Mdm2, which suppresses p53, and proteosomal degradation during hypoxia (139, 154). Conversely, during hypoxia, HIF1 can produce apoptosis via p53-dependent pathways and can provoke a G₁ arrest by upregulation of CDKIs (p21, p27) and subsequent hypophosphorylation of RBF (37, 62, 106, 163) (Fig. 3). The interaction of HIF1 with several signaling pathways has recently been discovered and will be discussed below.

View larger version (55K): [in this window] [in a new window]   Fig. 3.   Response of the hypoxia-inducible factor (HIF) system to hypoxic exposure. HIF1 is normally degraded during normoxia by the activity of prolyl hydroxylases (PHL), the von Hippel-Lindau factor (VHL), and the APC/C, thereby maintaining a low basal level of expression and activity. On stimulation by hypoxia, insulin-like growth factor (IGF), and other factors, the degradative activity of PHL and VHL is blocked. and HIF1 is stabilized at both the RNA and protein levels. Stable HIF1 is now able to dimerize with HIF1, bind to chromatin, and activate the transcription of several genes, including erythropoietin (Epo) and vascular endothelial growth factor (VEGF). I-NOS, inducible nitric oxide synthase; HO-1, heme oxygenase 1.

The spindle checkpoint. The spindle checkpoint is activated to inhibit cell cycle progression when mitotic spindles are disrupted or chromosomes have not attached to microtubules. The mechanism of inhibition involves the binding of components of the spindle checkpoint pathway (e.g., Mad and Bub proteins) to the ubiquitin-conjugating system and subsequent inhibition of the APC/C, which is normally involved in sister chromatid separation (Fig. 2).

The prophase checkpoint. A newly uncovered checkpoint, the prophase checkpoint, has been recently described. It is noteworthy that this required in vivo imaging to detect; however, this phenomenon has been reported previously in 1969 (17). Yet, this underscores the importance of the utility of the new technological advances that have been made in the recent past that reveal phenomena not previously amenable to investigation. A key facet of the prophase checkpoint is that cells that have initiated mitosis but have not progressed beyond prophase return to a premitotic stage, i.e., interphase, in response to chromosomal damage (61). Cyclin A/cdk2 appears to be the mediator of the prophase checkpoint, in line with its ability to be inhibited by p21. Reider and Cole (155) were also able to demonstrate that PtK1 cells in prophase returned to interphase if damage was induced <30 min before NEBD.

	THE DIRECT EFFECT OF HYPOXIA ON CELL CYCLE ACTIVITY

TOP ABSTRACT INTRODUCTION THE CELL CYCLE THE DIRECT EFFECT OF... SUMMARY REFERENCES

Animal Models

Over the past few decades, it has become apparent that several species, such as the turtle and brine shrimp, are capable of surviving extended periods of severe hypoxia or even anoxia (26, 88, 93). Several years ago, we discovered in our laboratory that D. melanogaster is also capable of surviving and recovering from several hours of prolonged and total O₂ deprivation (74, 123). Our discovery of the Drosophila O₂ tolerance is important for several reasons: 1) Drosophila is genetically very well studied and a large number of different fly lines and mutants are available for dissecting different signaling pathways; 2) it is now becoming clear that in almost all aspects of biology, from ion channels (94) to embryonic development and to neuronal development (5, 75), mammals and Drosophila use very similar complements of genes; and 3) with the completion of the sequencing of the entire Drosophila genome and now a significant portion of the human genome, the Drosophila becomes an increasingly useful model for deciphering the function of certain genetic pathways. It is therefore possible that by understanding the molecular mechanisms underlying the tolerance to anoxia in Drosophila, we will be able to better understand mechanisms of intolerance and injury in other species. More importantly, because studies have shown that tumor cells in humans can be very tolerant to anoxia (34), these studies raise the distinct possibility that cell manipulation can induce anoxia tolerance in human cells. This has recently panned out in some of our studies (Chen and Haddad, unpublished observations). Utilization of fruit flies as a model has already proven to be very useful in our studies in both adult flies (74, 122, 123) and embryos, in which we have addressed a number of very interesting and important questions.

Hypoxia induces variegated effects on early rapidly cycling embryos of different species and can affect the cell cycle in a variety of ways. The cell cycle can halt completely and irreversibly, leading to cell death. On the other hand, the cell cycle can halt its activity for a period of time during stress and can resume its activity when that stress is removed. Recovery from such a cessation of activity can be without a price or can compromise cell function in either the long term or the short term. There are many examples of cells that can transiently arrest cell cycle activity. Immortalized cells of human tumors and other mammalian cell lines have been demonstrated to arrest or delay cell cycle during a variety of stresses. For example, exposure to UV, IR, and spindle-disrupting drugs like nocodazole can lead to arrest (138). The length of the arrest and the sequelae that are injurious in the short and long term appear to depend on a multitude of factors, including the age of the organism, the species, the stage of development and the preexisting cell cycle status of the cell. In addition, the duration of the stress as well as the intensity of that stress also play important roles in the nature of the arrest and the extent of injury experienced. Within the cell cycle of both prokaryotes and eukaryotes are a number of transition states that can be interrupted if certain requirements have not been met or if the regulation of a number of cell cycle regulatory proteins has been altered (54, 82, 131, 156).

Brine shrimp, zebra fish, and nematodes. There appears to be specific stages within the cell cycle when a potentially survivable arrest can occur. Again, there is species variability, age, and stage dependence in this response. In studies of rapidly cycling embryos, Padilla et al. (142) recently reported that the nematode Caenorhabditis elegans can transiently enter a state of "suspended animation" or arrest during all stages of the cell cycle. C. elegans blastomeres arrest in interphase, prophase, metaphase, and telophase but not during anaphase. Zebrafish embryos (Danio rerio) also enter a state of reversible arrest (141); however, these embryos arrest during S phase and G₂ but not during mitosis. This indicates that there are species-specific hypoxia-induced checkpoints. Although the stages of cell cycle arrest were not addressed in these studies, it is remarkable that brine shrimp embryos (Artemia franciscana) can enter a quiescent state for several years and resume development on reoxygenation (27, 28, 77).

Fruit flies. In our particular paradigm, utilizing in vivo analysis of fruit fly embryos containing a GFP-kinesin construct that permitted real-time visualization of the response of fruit fly embryos to hypoxic and anoxic exposure, it was interesting to note that embryos arrested in only two positions in the cycle, i.e., at metaphase (with aligned, nonsegregated chromatids) and prior to the S phase (in G₁ or an hypoxia-induced interphase), as early cycling embryos do not possess a G₁ phase (42, 44) (Fig. 4). These arrest points depend on the stage of the cell cycle when the embryo was exposed to O₂ deprivation; i.e., interphase arrest occurs if hypoxia is induced after metaphase accompanied by some progression through the cell cycle and metaphase arrest is induced if hypoxia is applied shortly before metaphase and can be extremely rapid.

View larger version (59K): [in this window] [in a new window]   Fig. 4.   Effect of near anoxia on the cell cycle activity of Drosophila melanogaster embryos. In vivo confocal imaging of a green fluorescent protein (GFP)-kinesin embryo arrested at interphase during hypoxia: A: control, normoxic embryo, at the metaphase-to-anaphase transition of the cell cycle, captured 26 s before hypoxic perfusion. The embryo continues to cycle during hypoxia (B-E), passing through metaphase-anaphase (B), anaphase (C), and telophase-cytokinesis (D) before arresting in interphase (F). The embryo remains arrested for the duration of hypoxia (30-min hypoxic exposure) but demonstrates recovery (increased centrosomal fluorescence at 5 min of normoxia; G) and cell cycle activity (H) within 20 min of normoxia. Magnification = ×63. (Reprinted from Ref. 44, with permission.)

Role of Hypoxia in Tumor Cell Cycle Activity

The first studies of the direct effect of hypoxia on cell cycle activity were performed generally on transformed or oncogenic cells lines. The resistance of hypoxic cells in solid tumors to radiotherapy and chemotherapy had already been well documented (14, 170). It was noted that hypoxic cells in experimental tumors of varied origins could comprise from a few percent up to >80% of the viable tumor cell population (71). After irradiation, some tumors were able to reoxygenate, but this reoxygenation is variable in extent and time course (97). This led to the study of the effect of hypoxia on cell cycle activity and progression and cell viability in a variety of mammalian and human cell lines.

Shrieve and Begg (165) reported that rapidly induced, extreme hypoxia was able to immediately arrest V79-379A fibroblast cells in all phases of the cell cycle, similar to reports of Shrieve et al. (164) in EMT6/SF cells and by Petterson and Lindmo (149) in the human cell line NHIK3025. Utilizing bromodeoxyuridine (BrdU) incorporation and flow cytometry, Shrieve and Begg (165) were able to demonstrate that cells did not move into or out of S phase during hypoxia. Reoxygenation led to the initiation or completion of DNA synthesis; however, subsequent cell cycle progression was markedly delayed. Shrieve and Begg also reported that there appeared to be a slight progression of S-phase cells toward G₂. It was also noted that some early S-phase cells did continue to synthesize DNA during the first 2-4 h of hypoxia. During reaeration, cell cycle kinetics were apparently disturbed such that cell cycle progression was markedly slowed and some S phase cells demonstrated endocycling or overduplication of DNA. Similar reports were provided by Åmellem and Pettersen (3).

However, other reports indicated that cells have the capacity for continued growth in the nominal "absence of oxygen" (7, 11, 36, 158). Shrieve and Begg (165) believed that those differences were due to continuous O₂ contamination in the other studies. They concluded that a lesser degree of hypoxia or a more gradual induction of hypoxia could lead to different cell cycle activities and arrest points. Another observation of significance was that cell cycle perturbations that occurred during reoxygenation such as cell cycle slowing and endoreplication were dependent on the phase of the cell cycle that cells were in during hypoxia, in agreement with the studies of Petterson and Lindmo (149) and Shrieve et al. (164). The value of these previously mentioned studies was the revelation that cells can delay or arrest cell cycle activity during near anoxia and can continue cycling on reoxygenation. Furthermore, the stage of the cell cycle at the time of hypoxia induction may influence the stage at which the cell arrests.

Several questions remain to be answered, and new questions were generated by these early studies in cancerous cell lines. For example, why and how did cells survive near anoxia, and why did these cells not undergo apoptosis? Some of the answers had to wait for the molecular biology revolution that occurred in the 1980s and 1990s to be answered. For instance, it was realized that several cancers and many tumors cells did not have a functional p53 protein or RBF protein and were therefore unable to induce p53-mediated apoptosis or a RBF-induced G₁ arrest (99, 106). This led to a revolution in the understanding of regulation of the cell cycle machinery during normal development and dysplasia and neoplasia.

Hypoxia, Arrest of Cell Cycle, and Various Mechanisms

There have been a number of potential mechanisms of arrest during stressor hypoxia, with some identified in mammalian cell lines and some in invertebrates. First, O'Farrell and colleagues (42) recently showed that ATP levels and nitric oxide play a role in syncytial embryos, which not only arrest but actually die when they are subjected to hypoxia, in contrast to cellularized embryos that arrest and survive but in which nitric oxide and ATP are not major players. Second, studies have shown that ascites tumor cell line exposed to hypoxia do not progress through the G₁ phase and do not synthesize DNA because of depletion of deoxynucleotides from an inhibition of enzymes, such as ribonucleotide reductase and dihydroorotate dehydrogenase, responsible for the synthesis of pyrimidines and DNA bases (10, 116, 117). Most likely, this mechanism cannot play a role in organisms such as D. melanogaster embryos, which arrest within seconds of instituting hypoxia, since there are usually pools of purines and pyrimidines from which DNA can be synthesized. Third, other studies have demonstrated that the initiation or replication of DNA may be sensitive to hypoxia and that this is the reason for the cell cycle arrest (151). Fourth, more specific mechanisms include, for example, the effect of hypoxia on RBF phosphorylation and whether the hypophosphorylation state of RBF found in hypoxia (106, 120) could be a mechanism by which the arrest during G₁ takes place. The effect of hypoxia on the rest of the important proteins that control the G₁/S transition, which include E2F, Dp and cyclins D and E, is ill defined and not well studied. Furthermore, DNA damage checkpoint proteins may play an important role in the G₁/S arrest. If there is any "damage" to DNA during hypoxia or ischemia (which in some cell lines may take place not over minutes but hours), cells may activate mechanisms that would protect them from defects in heritability. Hence, sensor proteins like poly-ADP-ribose polymerase, Rad1, Rad9, Hus1, or Rad17 and their homologs may be important for hypoxia studies. Similarly, kinases such as ATM and ATR, proteins such as Chk1 and Chk2, and downstream effectors (cdc25 and cdks) may be crucial in the study of the overall sensing and transduction of hypoxia into an arrest (Figs. 1 and 2).

Other potential mechanisms for cell cycle arrest have been studied in our laboratory. Our laboratory has shown (44) that the cell cycle can arrest in hypoxia not only in the pre-S interphase but also in metaphase; therefore, proteins such as the securins and separins, which control sister chromatid cohesion and separation as well as a block on the APC, could be involved and perhaps critical in this hypoxia-induced metaphase arrest (90, 104). Our laboratory (44) has also demonstrated that cyclin B is not degraded during a hypoxia-induced metaphase arrest; therefore, it can be hypothesized that O₂ deprivation imposes a block on cyclin B degradation either by maintaining the inactivated state of the APC or by preventing activation of the APC.

Metabolic theories also exist. Although these do not pinpoint the exact location of the arrest in the cell cycle, cell division is involved. For example, fructose 2,6-bisphosphatase when overexpressed decreases glycolysis and delays the progression of the cell cycle. It is interesting that this effect of the fructose 2,6-bisphosphatase was mostly related to its kinase domain that was being upregulated alone in cells in vitro (148). Other studies have investigated the effect of hypoxia on insulin-like growth factor (IGF) binding protein and how this may affect cell growth and division in the fetus in utero or in cells in vitro (Fig. 3) (150, 169). Also, recent studies showed that hypoxia inactivates glycogen synthase kinase via phosphoinositol 3 (PI3)-Akt pathway (20), which can impact cell division and proliferation. For instance, previous studies suggested that the activation of this Akt pathway modulates cyclin D turnover, increasing its stability and allowing cells to progress through the G₁/S transition (145). Giaccia and colleagues (186) also recently linked hypoxia to phosphatase and tensin homology deleted on chromosome 10 (PTEN), which functions as a tumor suppressor, and HIF1 and found that cells lacking PTEN enhances HIF1 expression and the rescue of PTEN in this same cell line ablates the hypoxia-induced HIF and downstream gene expression (186).

In addition to the mechanisms described or alluded to above, hypoxia also seems to have specific effects on specific cell types. Laderoute and colleagues (109) demonstrated recently that hypoxia can have opposing effects on the expression of certain genes and proteins in different cell types. For example, although hypoxia induces vascular endothelial growth factor in all human cell types tested, hypoxia induced a decrease in thrombospondin 1, an inhibitor of angiogenesis in specific tumor tissues. Furthermore, although tumors can induce disturbed coagulability for many years, it is only now that we have been able to learn about the mechanisms that induce these defects. It turns out that hypoxic tumor cells release substances that affect coagulability. For example, it is now well known that plasminogen activator inhibitor, uPA, uPAR, angiogenin, angiopoeitin and platelet-derived growth factor are only some of those proteins and substances shown to be hypoxia induced and can affect coagulation (for review, see Refs. 40 and 103). Another very interesting effect of hypoxia is on adipogenesis. Fascinating results have been obtained on fibroblasts deficient in HIF1. It was found that hypoxia inhibits uPAR nuclear transcription, which in turn limits adipogenesis. HIF^/ cells do not respond to hypoxia-induced inhibition because a downstream gene to HIF, DEC1/Stra13, does not get activated by HIF and does not repress uPAR to inhibit adipogenesis. Hence, hypoxia inhibits adipocyte differentiation. Knowing the mechanisms that can lead to obesity may allow us to control this disease by potentially mimicking the effect of low O₂.

It should be clear from the preceding remarks that there are a number of genes and proteins that control the hypoxia-induced changes in cell and tissue phenotype. For example, a number of genes determine the influence of hypoxia on cell proliferation or cell cycle arrest and apoptosis. How the mitochondrial-signaling pathway of Apaf/caspase 9 leads to cell death via caspase 3 is involved in signaling of the cell cycle is not clear at present especially when it is related to hypoxia and cytochrome release. There is evidence, however, that this Apaf-pathway is independent of p53 and bax (2).

Although many studies have been performed to investigate the hypoxia-induced cell cycle arrest, only a few studies have determined some of the pathways responsible for reentry into the cell cycle after such an arrest. It seems from these studies that such mechanisms are not similar necessarily to those that induce cell cycle arrest. p21 and p27, for example, do not play a major role in the cell cycle arrest but in the cell cycle reentry. Indeed, cells lacking p21 resume DNA synthesis more rapidly with reoxygenation than p21^+/+ cells. p27 may have also a similar role in applying "brakes" in cell cycle reentry.

One of the exciting and interesting questions in this area is the effect of hypoxia on development in very early life. Although many investigators have posed similar questions in the past, we have started to appreciate how important is the titration of O₂ in early life for proper development of organs and body structure. Despite the fact that fetuses are naturally hypoxic, further hypoxia or oxygenating fetuses above their usual level of O₂ can induce developmental defects (19). How development takes place at high altitude, what are the effects of various altitudes, and what are the functional and structural consequences at high altitude are not well studied at present. Some of the corollary issues in this regard is the effect of hypoxia on cell size vs. cell growth. It is clear now, as has been commented on above, that a number of genes are important in the control of cell size and cell cycle timing and growth. For example, overexpression of E2F or RBF does not increase the size of the organ but modulates the number of cells and therefore cell size in that particular organ (134). The insulin-IGF-PI3 kinase pathway in both vertebrates and flies plays a major role in the control of cell size (30). How that pathway is involved in the hypoxia-induced growth changes that are both clinically and experimentally observed is not well studied at this stage. In addition, it is not clear whether the nutrient-activated, rapamycin-sensitive protein kinase pathway is involved during hypoxia and how that affects cell growth and size. Other questions regarding development and adaptation at high altitude across multiple generations have not been reported.

Hypoxia, as a stressful stimulus, has taken a different meaning in the past decade. Because of the nature of the research done in hypoxia on so many different types of cells, in vertebrates and invertebrate models, and in so many different aspects of cell biology and physiology not only in healthy tissues but also in diseased or mutated cells, we have learned how to take advantage of the effects of hypoxia to treat diseases. For example, the scientific community has taken advantage of what hypoxia does to tumor blood vessels to combat tumors via antiangiogenic drugs. Because we now know that tumors have hypoxic cores, Liu and colleagues (115) genetically engineered anaerobic bacteria that target hypoxic regions of solid tumors. These bacteria carry enzymes that transform into cytotoxic substances when bacteria, the vectors, reach a hypoxic environment. It will be fascinating and certainly important to exploit further our understanding of what genes hypoxia induces to better treat human diseases.

	SUMMARY

TOP ABSTRACT INTRODUCTION THE CELL CYCLE THE DIRECT EFFECT OF... SUMMARY REFERENCES

Hypoxia induces many effects on cells and tissues. In this review, we have briefly outlined some of the salient features of the effect of hypoxia on cell cycle in both vertebrates and invertebrates. We have also highlighted the importance of D. melanogaster as an important model in trying to understand the effects of hypoxia on cell division, growth, and proliferation. Our hope is not only to stimulate interest in this area of investigation but also to show that the understanding of disease at the molecular level for use in new therapeutic modalities presents an exciting future.