Vascular endothelial growth factor signaling in hypoxia and inflammation.

Inflammation refers to physical and physiological changes observed at the site of injury or infection. The route word for inflammation in Latin is ‘inflammo’ meaning ‘ignite’. Increased redness, heat, pain, and swelling are associated with inflamed tissues. Redness is caused by vasodilation accompanied by vascular stasis causing reduced movement of blood cells. Vasodialation is casued by bradykinin, histamine and VEGF released at the site of tissue damage and nitric oxide generated by vascular endothelium. Stasis is a culmination of changes occurring at the vasculature. Histamine and VEGF induce vascular leak at the site of inflammation leading to exudation of plasma. Vascular permeability increases cell density inside the vessels and interstitial pressure outside the blood vessels. Interstitial pressure can collapse capillaries and increase thrombotic events. Resulting sluggish blood flow causes stasis, which inadvertently helps the innate immune cells to extravasate into tissues. Endothelial cells respond to inflammatory signal by upregulating cell adhesion molecules such as ICAM, P-selectin and E-selectin. E-selectin, CD62, is a cell adhesion molecule selectively upregulated in endothelial cells upon cytokine stimulation. E-selectin has a lectin binding domains at the N and C termini, EGF like domain in the middle along with six-suchi domain repeats (CCR, complement control protein). E-selectin binds to sialyl-Lewis x carbohydrate antigen present on neutrophils, monocytes, eosinophils, T-cells and NK-cells. ESL-1 is described as a counter-receptor for E-selectin and is expressed on inflammatory cells. Expression of E-selectin is induced by P-selectin, IL-1β and TNF-α by transcriptional upregulation ( 108 – 111 ). P-selectin, CD62P, another cell adhesion molecule that is expressed on endothelial cells and platelets is an early responder to inflammatory stimuli ( 112 ). P-selectin is stored in vesicles and are quickly expressed on cell surface by vesicle fusion (degranulation). Platelets and endothelial cells respond to thrombin, type II collagen, ATP and respond by expressing P-selectin on cell surface. Similar to E-selectin, P-selectin binds to PSGL-1 ligand expressed on inflammatory cells ( 113 – 115 ). Expression of E and P-selectins are modulated by sheer stress of blood flow and cytokines. Hypoxia alone is not sufficient for E-selctin upregulation but it is superinduced in the presence of cytokines such as TNF-α or LPS. Intercellular adhesion molecule, ICAM1, and VCAM are other important mediators of inflammatory cell attachment to endothelial cell under hypoxic conditions. ICAM1 (CD54) is a glycoprotein expressed on endothelial cells, macrophages and lymphocytes when stimulated with TNF-α or IL-1α. ICAM binds to LFA-1, an integrin (αL, CD11α; β2, CD18) expressed on lymphocytes and macrophage adhesion ligand-1, MAC-1, ( 116 , 117 ). ICAM1 can also be induced by RANTES secreted by macrophages and granulocytes. ICAM1 is highjacked by rhinovirus ( 118 ) and malaria parasite infected erythrocytes for transendothelial cell migration ( 119 ). Recent studies suggest that ICAM1 upregulation during hypoxic stress is independent of HIF-1 but dependent on NFκB-mediated transactivation ( 120 ). Even though ICAM1 expression both on macrophages and endothelial cells were independent of HIF-1 but mediated by a novel pathway regulated by prolyl-hydroxylases (PHD) during hypoxia. In a recent study, the role of PHD isoforms on inflammatory response was elucidated in gene knockout models ( 121 ). These studies showed that PHD3 is very important for controlled innate immune response during sepsis. PHD3 deficiency (PHD3−/−) aggravated innate immune response leading to early organ failure and death in a HIF-1 and NFκB dependent manner. VCAM-1 (CD106) is an another endothelial cell adhesion molecule upregulated during inflammation and hypoxia ( 122 ). VCAM-1 binds to VLA-1 antigen (alpha4 beta1 and alpha4 beta7 integrins) expressed on lymphocytes. VCAM-1 also interact with erzin and Moesin, which act as intermediates of cell membrane (ICAM-2) to actin bundles thereby regulating cell attachment and migration. Blood borne monocytes and lymphocytes are recruited to inflammatory sites by a complex signaling cascade that involves a plethora of cytokine and chemokines. Circulating monocytes for example attach to cell adhesion molecules over expressed on endothelial cells under hypoxic/inflammatory stimuli (rolling model). Subsequently, monocytes extravasate to reach tissues by a processes called diapesis, squeezing between endothelial cells without affecting barrier functions. Outside the vasculature monocytes differentiate into tissue macrophage or dendritic cells in mediating innate or adaptive immune response respectively. VEGF, CXCL12, endothelial monocyte activating peptide-II (EMAPII) and Angiopoietin-2 (Ang2) mediate extravasation of monocytes under hypoxia. Upon extravasation, human monocyte derived macrophages (hMDM) adapt to hypoxic environment quickly by differentially expressing genes such as FGF, VEGFR and hypoxia-inducible proinflammatory cytokines such as IL-1beta, TNF-alpha, and acute phase protein, IL-6 ( 123 ). A comprehensive list of chemokines elaborated under hypoxia is summarized in Table 1 and 2 . In addition to chemokines, chemokine receptors are also modulated in monocyte-derived macrophages at hypoxic environment. Of particular interest is that hypoxia inhibits migration of macrophages by inhibiting cell motility promoting factor, and inducing the expression of GRO family of chemokines such as CXCL2 and CXCL3 which are negative regulators of macrophage migration ( 124 ). Thus, hypoxia traps monocyte-derived macrophages after their extravasation without affecting their ability to differentiate into immature dendritic cells (iDC). Immature dendritic cells are highly migratory and reach the draining lymph nodes. iDC differentiate to mature CD (mDC), which are retained at the site of inflammation. Hypoxia at the site of inflammation therefore modulates chemokine and chemokine receptor expression in inflammatory cells. One of the final outcomes of these changes is recruitment of monocytes to local site of tissue damage, differentiation of monocytes to tissue macrophage, retention of macrophages at hypoxic site and generation of highly motile immature dendritic cells and less motile mature dendritic cells. These changes are achieved by altering transciptome of monocyte/macrophage by hypoxia. For example, nuclear factor-kappa b (NFκB), the central transcription factor mediating inflammatory response is intricately involved in hypoxia-induced expression of monocyte chemoattractant protein-1 (MCP-1). VEGF induces MCP-1 through the activation of NFκB pathway ( 125 ). MCP-1, in turn, can induce VEGF thereby establishing a feed-forward loop in human aortic endothelial cells ( 126 ). Inflammatory cytokine, IL-1β, induced NFκB-cyclooxygenase-2 axis was found to stabilize HIF-1 and as a consequence induced VEGF ( 127 ). Similarly, TNF-α-induced TLR4-NFκB pathway intersects with HIF-1α mediated inflammatory response as well ( 128 ). These studies reiterate cross talk between inflammatory cytokines and hypoxia-induced responses in recruiting and regulating innate immune cells. VEGF is effective in treating ischemia-induced damage of neuronal tissues ( 129 ). VEGF administration induced not only angiogenesis but also protected neurons from cell death and neurogenesis in preclinical models of focal cerebral ischemia ( 83 ). VEGF gene therapy using AAV has a significant impact on reducing neuronal damage following stroke ( 129 ). VEGF reduced inflammatory cytokine levels in the brain following ischemic stroke and attenuated immune cell infiltration ( 130 ). An alternate approach is to transplant CNS stem cells, which secrete significant amounts of VEGF. Cell-derived VEGF was necessary to recover from ischemic injury. VEGF was necessary to establish BBB and suppress inflammation in addition to neovascularization. Such an approach has been successful in promoting angiogenesis and tissue remodeling at the damaged site ( 131 ). VEGF therapy increases pericyte covering, vessel normalization and improved blood flow in experimental animals ( 132 ). Same group of researchers have again showed that VEGF mediated signaling and recovery following stroke was attenuated in hyperlipidemia or in ApoE−/− genetic background ( 133 ). This raises the potential limitation of VEGF centric therapeutic strategies to improve angiogenesis and blood flow to limit stroke-induced tissue damage. Various in-situ hybridization studies have shown increased VEGF mRNA levels in various solid tumors including carcinomas of lung, gastrointestinal tract, breast, endometrium, urogenital tract and intracranial tumors ( 134 ). The expression is particularly found to be correlated to hypoxia in the tumor cells. Although, tumor cells are the major source of VEGF, stromal cells have been identified to be an important source as well ( 135 ). Tumor stroma is made of fibroblasts, inflammatory cells, adipocytes and endothelial cells. This unique microenvironment resembles tissues under persistent inflammation in many respects. Tumor hypoxia and inflammation are prognostic in several types of cancers. Higher levels of VEGF secreted by tumor cells and stromal components helps in further mobilization of mesenchymal stem cells and precursors of inflammatory cells from the bone marrow. VEGF induces tumor angiogenesis. Consequently, VEGF inhibitors are effective in inhibiting tumor growth and metastasis. Blocking VEGF and its receptors has been much more effective than monotherapy. Gerber et al showed that chimeric receptor containing first three Ig-like domains of VEGFR-1, common to both human and mouse, resulted in nearly complete suppression of tumor growth in a mouse model of human rhabdomyosacroma ( 136 ). VEGF inhibitors combined with chemotherapy and radiotherapy resulted in better tumor suppression compared to either therapy alone ( 137 ) ( 138 ). Klement et al found that combining a monoclonal neutralizing antibody (DC101) targeting the flk-1/KDR (VEGFR-2) with low dose vinblastine chemotherapy enhanced the anti-vascular effects ( 139 ). Currently, many VEGF inhibitors are in various phases of clinical development (antibodies to VEGF, VEGF-R, VEGF-neutralizing aptamers and inhibitors of receptor tyrosine kinases). One of the strategies to inhibit VEGF-VEGFR signaling is a humanized antibody against VEGF, Avastin (Bevacizumab) which has been approved by FDA for clinical use. Similar to the solid tumors, VEGF is also expressed in a variety of hematological malignancies like T and B-cell lymphomas, acute and chronic myeloid leukemia, multiple myeloma and Burkitt's lymphoma ( 140 ). VEGFR-1 and VEGFR-2 are also detected in some leukemia cell lines, however VEGFR-2 is found more frequently compared to VEGFR-1 ( 141 ). Additionally, a neutralizing monoclonal antibody IMC-1C11, specific to human VEGFR-2 inhibited proliferation of xenotransplanted human leukemia cells and increased survival in the mouse models ( 142 ). Based on the promising results in pre-clinical trials, many of the VEGF inhibitors are currently being tested in the clinical trials of hematological malignancies. Polycystic ovarian syndrome (PCOS) is an important cause of female infertility characterized by hirsutism, obesity, polycystic ovaries and menstrual irregularities. Hyperplasia of stroma and theca contributes to the excessive androgen production responsible for the symptoms of PCOS. Stromal angiogenesis is an important feature of the polycystic ovaries. VEGF levels are elevated in PCOS patients, compared to healthy females ( 143 ). VEGF mRNA levels were also elevated in the cyst walls. An increased expression of VEGF is associated with increased ovarian stromal blood flow, which is believed to disrupt the ovarian autoregulatory mechanisms, thereby leading to uninhibited growth of all cohort follicles. These findings reflect that VEGF may be an important player in the pathogenesis of PCOS. Another important condition characterized by pathologic angiogenesis is endometriosis. Endometriosis is a condition, in which endometrial implants are found outside the uterine cavity. These ectopic implants develop vessels that enable them to survive and grow. Increased VEGF levels are observed in the peritoneal cavity of endometriosis patients ( 144 ). Two VEGF inhibitors, a soluble truncated receptor (decoy) and an antibody to VEGF were recently studied in the mouse models and were found to result in significant reduction in the endometrial implant size ( 145 , 146 ). VEGF is implicated in preeclampsia, a serious obstetrics complication leading to hypertension, proteinuria, and glomerular endotheliosis. Increased circulating levels of soluble VEGFR1 (sFlt1) is observed in pre-eclampsia patients ( 147 ). This study demonstrated that endothelial dysfunction was a result of decreased VEGF and PlGF as a consequence of increased levels of sFlt1. Exogenous administration of VEGF and PlGF reversed pre-eclampsia in model systems. Current studies are in progress to evaluate the use of VEGF therapy in pre-eclampsia patients. The conditions associated with intraocular ischemia include diabetes, central retinal vein occlusion, prematurity and wet-type of age-related macular degeneration. These conditions subsequently lead to intraocular neovascularization which is associated with grave consequences including vitreous hemorrhages, retinal detachment and blindness ( 148 ), ( 149 ). Earlier studies have shown increase in VEGF in aqueous and vitreous humor associated with intraocular ischemia, contributing to the neovascular disorders. Various pre-clinical and clinical studies are currently investigating the effects of VEGF inhibitors in these conditions ( 150 ). The VISION trial first showed that pegaptanib, an anti-VEGF agent was able to prevent vision loss in neovascular age related macular degeneration. Fab fragment of a humanized anti-VEGF antibody, Lucentis (Ranibizumab) is now clinically used to treat wet, age-related macular degeneration. The ability of VEGF to regulate pathologic angiogenesis in ischemic conditions, led to prospects in vascular occlusive diseases, in which restoration of blood flow is life saving. VEGF targeted therapies have been tried in myocardial infarction, peripheral limb ischemia and stroke ( 151 ). An increase in vascularity was observed with adenoviral and plasmid liposome mediated delivery of VEGF 165 in limb ischemia patients ( 152 ). A double blind randomized controlled trial including diabetic patients with peripheral arterial disease also showed a significant improvement in symptoms in patients treated with intramuscular VEGF 165 ( 153 ). Patients treated with VEGF had fewer amputations and more skin ulcer healing compared to the placebo group. RAVE, a double-blind trial assessing the intramuscular adenoviral gene transfer of VEGF 121 however showed no improvement in the symptoms, in patients with limb ischemia ( 154 ). This raises important concerns regarding the difference in efficacy of the various isoforms of VEGF and the importance of consistent method to measure primary endpoint of improved perfusion in different trials. A phase II KAT trial, assessing the efficacy of VEGF 165 gene therapy in patients with coronary artery disease also found improvement in myocardial perfusion, but without any change in rate of restenosis ( 155 ). Another trial, Euroinject One, showed significant improvement in the ventricular function in patients treated by intramyocardial VEGF 165 plasmid gene transfer ( 156 ). However, the results in myocardial infarction have been disappointing in terms of improvement in myocardial stress perfusion and angina class. In an effort to improve these outcomes, Ripa et al conducted a pilot study of combined VEGF 165 gene therapy and stem cell mobilization ( 157 ). There was still no improvement in primary end point of myocardial stress perfusion probably because of inadequate homing. Future trials should include co-transfer of plasmid encoding SDF-1, a homing factor, timely administration of G-CSF and a consistent method for perfusion assessment. Overall, current evidence suggests numerous factors that could result in lack of benefit in some efficacy end points. Several laboratories are currently investigating these factors and the effects of targeted VEGF therapies in various ischemic diseases. Hypoxia-induced changes at the transcriptome and inflammation-induced cytokine/chemokine responses modulate endothelial cells and innate immune cells. As cancer, cardiovascular and infectious diseases are modulated by hypoxia and inflammation, it provides an opportunity to develop novel drugs to intervene this co-dependent signaling pathways. Hypoxia inducible factor, HIF, is an evolutionarily conserved key transcription factor that regulates vascular biology, angiogenesis, metabolic reprogamming and inflammation. HIF is regulated at the cellular level by canonical pathway involving proteasomal degradation and noncanonical regulation involving HSP-70, NFκB and microRNA. HIF regulates recruitment of inflammatory cells, erythropoiesis, tissue remodeling, and pH homeostasis. Furthermore, HIF-mediated metabolic adaptation creates a niche for stem cells to thrive in a hypoxic microenvironment. One of the primary target genes regulated by HIF is VEGF-A. VEGF-family of growth factors is necessary for developmental angiogenesis, collateral vessel growth in peripheral arterial disease, stroke, and myocardial infraction. Inducing local production of VEGF is therapeutically useful in limiting hypoxia/ischemia-induced tissue damage. Conversely, pathological angiogenesis can be inhibited by strategies to neutralize VEGF or inhibit its signaling pathways. VEGF family of growth factors bind to three distinct receptors. Pleotropism in receptor binding and differences in tissue distribution of receptor/co-receptor offers therapeutic strategies to modulate VEGF signaling. VEGF-induced changes in inflammatory signals and its positive influence on neuroprotection provide a road map for the future development of VEGF-based therapies.

Ischemic diseases are the major cause of morbidity and mortality worldwide. The spectrum of ischemic diseases is wide and includes both acute and chronic conditions. Ischemia can manifest as angina, acute coronary syndrome and chronic ischemic heart disease; as stroke in brain; as acute or chronic mesenteric ischemia in gastrointestinal system and as peripheral vascular diseases in limbs. Ischemic diseases share common risk factors, pathophysiology and etiology ( 1 , 2 ). It is well known that ischemia initiates an intense inflammatory response that is clinically relevant and worsens the ischemic injury ( 3 , 4 ). In stroke, cerebral ischemia is accompanied by infiltration of inflammatory cells, which is initiated by ischemia-induced expression of cytokines, adhesion molecules and other inflammatory mediators. Infiltration of inflammatory cells can lead to activation of microglia and astrocytes and persists for hours to days. Inflammation in ischemic tissues exacerbates tissue damage leading to worse neurological outcomes ( 5 – 8 ). The inflammatory changes also occur in ischemic myocardium and peripheral vascular diseases. Ischemic myocardium triggers activation of complement system leading to monocyte and neutrophil recruitment ( 9 , 10 ). Atherogenesis, an important factor in the pathogenesis of cardiovascular and limb ischemia, is also linked to inflammation ( 11 , 12 ). Furthermore, obesity is another clinical condition, in which hypoxia is associated with inflammation. The enlarging adipocytes experience an imbalance between demand and supply of oxygen leading to an increase in the secretion of inflammatory adipokines in the fat tissues. Chronic, low-grade inflammation in adipose tissues results in insulin resistance ( 13 ). As hypoxia induces inflammation, the reverse is also true in certain other diseases, wherein, inflammation causes hypoxia. Colitis ( 14 ), acute lung injury, trauma and infection are some of the examples. Inflammation impairs tissue perfusion because of thrombosis. Vascular leak leads to edema, increased interstitial fluid pressure and compression of vessels (colitis), atelectasis of airways (acute lung injury, chronic obstructive pulmonary disease) and trauma ( 15 ). Thus, cellular responses to hypoxia is central to the pathophysiology of major diseases ( Fig.1 ). Hypoxia inducible factor (HIF) is a pivotal transcription factor induced under hypoxia which transactivates target genes such as vascular endothelial growth factor (VEGF). In this review we will focus on the molecular regulation of HIF and VEGF signaling in inflammation and ischemia.

There are three members in the family of HIF, HIF-1, HIF-2 and HIF-3. HIF-1 is the founding member. HIF-2 was discovered later ( 31 , 32 ). HIFs are heterodimeric proteins consisting of an alpha subunit and a beta subunit. Alpha subunit is induced during hypoxia while the beta subunit is constitutively expressed. Beta subunit is called aryl hydrocarbon receptor nuclear translocator (ARNT). HIF belongs to the PER-ARNT-SIM (PAS) subfamily of basic helix-loop-helix (bHLH) family of transcription factors. Both alpha and beta subnits have similar structural domains. NH-2 domain contains a bHLH domain that mediates DNA binding. Central region has two PAS domains, PAS-A and PAS-B, which facilitate heterodimerization and a Cterminal region that regulates recruitment of co-activators and co-repressors (TAD, transactivation domain). TAD binds to co-activators such as CBP/p300, histone acetyltransferases/deacetylases and target gene specific factors such as Smad3 and C/EBP alpha. Most important functional domain of HIF-1, 2alpha is the oxygen dependent degradation domain (ODDD), which determines the stability of the protein under different oxygen concentrations. HIF-3a is less studied and splice variants of HIF-3a have been suggested to play a dominant negative role in regulating HIF-1-mediated transactivation. HIFs are vital for developmental angiogenesis. Deletion of HIF-1 or HIF-2α leads to embryonic lethality at E10.5 stage due to defects in cardiovascular development ( 33 – 35 ). Conditional knock out mice were then generated to investigate the critical role of HIFs in developmental and pathological angiogenesis. Intracellular HIF levels and its localization are dependent on oxygen levels. HIF-1α will be considered as a prototypical example in this review. In normoxia, HIF-1α undergoes posttranscriptional modifications in the ODDD region. Prolyl hydroxylases (PHD1,2,3) are key oxygen sensing molecular switches. PHDs are a family of 2-alphaketoglutarate-dependent dioxygenases. PHDs use ascorbic acid, iron and oxygen as cofactors in modifying prolyl residues in the presence of 2-alphaketoglutarate (oxoglutarate). PHDs hydroxylate two proline residues located at position 402 and 564. Hydroxylated HIF-1α is then recognized by von Hippel Lindau (VHL) protein and assembled into E3 ligase complex consisting of VHL, Cullin-2, Elongin B/C and ring box protein 1 (RBX1), VCBCR complex. Subsequent polyubiquitation of HIF-1α targets it to 26S proteasome for degradation (canonical pathway, Fig. 2 ). This process is very efficient and as a result very low levels of HIF-1α are maintained under normoxia. Under hypoxia however, PHDs cannot function efficiently due to lack of oxygen, a cofactor, and HIF-1α hydroxylation is attenuated. Thus, HIF-1α degradation is prevented in low oxygen concentration. HIF-1α then heterodimerize with HIF-1β and translocate to the nucleus. Though PHD-dependent hydroxylation is the predominant mechanism of HIF-1α regulation, another posttranslational modification may also play a role in HIF-1α stability. A lysine residue at position 532 of ODD domain is acetylated by acetyl-transferase arrest defective-1 (ARD1) enzyme in an oxygen dependent manner. Acetylated HIF-1α binds to VCBCR complex efficiently and thereby results in enhanced degradation of HIF-1α. Unlike PHD enzymes, oxygen levels do not directly affect the functional activity of ARD1 but ARD1 levels change under hypoxia by transcriptional regulation. Role of ARD1 in HIF-1α regulation however remains controversial. Another important regulatory pathway that impinges upon the functional activity of HIF-1α is, hydroxylation of an asparagine residue at position 803. Factor inhibiting HIF-1, FIH, mediates the hydroxylation of Asn803 and prevents HIF-1 complex interaction with co-activators, CBP/P300. Inability to interact with co-activators attenuates the ability of HIF-1 to transactivate target genes. FIH activity is directly affected by oxygen, iron and oxoglutarate, which is very similar to the requirements for PHD enzymes. Additionally, HIF-1α is also phosphorylated by mitogen-activated protein kinase (MAPK) p42/p44 and p38, which increases stability and translocation into the nucleus. Other post-translational modifications such as S-nitrosylation increases the transactivating function of HIF-1 and SUMOylation on the other hand represses HIF-1 function. Furthermore, chaperone proteins HSP70 and HSP90, which are stress-induced protective factors, can sequester HIF-1α and prevent its degradation (non-canonical regulation). Chaperone-mediated autophagy seems to affect HIF-1 stability whereas macroautophagy did not influence HIF-1 stability ( 36 ). In addition, sirtuins, NAD+-dependent histone deacetylases, influence stability of HIF. Sirt1 was found to differentially deacetylate HIF-2α and not HIF-1α ( 37 , 38 ). However, another study found that Sirt1 could stabilize HIF-1α ( 39 ). Thus role of Sirt1 in HIF-1 activity remains inconclusive and contextual. A recent study by Semenza’s group showed that Sirt7 physically interacts with HIF-1α and HIF-2α and negatively regulates their levels. Sirt7 mediated effects on HIF were found to be independent of the enzymatic activity ( 40 ). Thus, complex regulatory pathways fine-tune the stability and function of HIF-1 under hypoxic stress. HIF-2α is a structurally and functionally related transcription factor. HIF-2α is also referred as EPAS1 and contains structural domain organizations similar to HIF-1α, Basic HLH, PAS and TAD domains. It shares about 48 % homology in primary sequence with HIF-1α. HIF-2α is highly expressed in some tissues and cells such as lung vasculature and carotid body. HIF-1α and HIF-2α transactivate genes that are common to both as well as distinct groups of genes ( 41 , 42 ). For example expression of IL-8 and Myc in endothelial cells are dependent on HIF-2α and not HIF-1α ( 43 ). HIF-2α is a major transcription factor expressed in endothelial cells and seems to play a major role in pathological angiogenesis ( 44 ). HIF-2α is also the primary driver for erythropoietin expression. In some experimental conditions, loss of HIF-1α is compensated by an increase in HIF-2α. Current evidence suggest tissue specific role for HIF-1α and HIF-2α in hypoxic responses. Unlike HIF-1α and HIF2-α, HIF-3α has not been extensively investigated. HIF-3α has multiple splice variants and expressed in tissue specific manner. Splice variants, HIF-3α2 and HIF-3α4 are shown to act as dominant-negative regulator of HIF-1α function ( 45 , 46 ). As HIF-3α suppresses transcriptional activity of HIF, it is also called Inhibitory PAS domain protein, IPAS. HIF-3α has been found to interact with Bcl-xL, and may induce proapoptotic signaling ( 47 ). IPAS is transcriptionally upregulated by HIF-1 suggesting a negative feedback loop regulating HIF-1-mediated target gene expression ( 48 ). MicroRNAs (miR) are short non-coding RNA 21–23 nucleotide long arising from either introns or exons. There are around one thousand miR characterized in human genome. MicroRNAs are transcribed as a longer (about 70 nt long) strand of RNA with unique stem loop structure (primary miR). Pri-mRNA is transported to the cytoplasm by exportin 5. Inside the cytoplasm they mature into pre-miR and then processed into a double stranded short segment of about 21–23 nucleotide by Dicer. Ago2 binding later selects either one of strand (3p or 5p) to base pair with target complementary sequence in 3’ untranslated region (3’UTR) of mRNA. The first 7 or 8 nucleotides of mature miR are termed ‘seed’ sequence. Binding of seed sequence at target sites in mRNA is necessary for regulating the function of transcripts. A complete match with target sequence results in degradation of target transcript while, an incomplete match affects translational initiation complex by interfering with Eukaryotic translation initiation factor 4G (eIF4G). Net result is inhibition of target protein expression. Each miR can target hundreds of targets and each target mRNA can have target sequence for multiple miR. Hypoxia in inflammatory sites can upregulate or down regulate classes of miR which in turn will have a reciprocal effect on target protein expression. Sequestering miR by competing transcripts can also modulate effects of miR. Our studies have established that under hypoxia, miR-424 is over expressed in endothelial cells. Increase in miR-424 was mediated by PU.1 transcription factor, which was regulated by C/EBP and RUNX2. miR-424 targeted CUL-2, the scaffolding protein which destabilized the E3 ubiquitin complex assembly and prevented degradation of HIF-1/HIF-2α ( 49 , 50 ). Hypoxia also down regulates the levels of miR-199a-5P, which targets HIF-1α ( 51 , 52 ). Reduction in miR-199a-5P contributes to increase in HIF-1α transcripts and protein. Thus, a web of miR network fine-tunes the stability of HIF-1 under hypoxia. Furthermore, increased HIF-1α under hypoxia transactivates the expression of miR-210 by binding to HRE elements present in the promoter of miR-210 ( 53 – 55 ). miR-210 is involved in suppressing mitochondrial function by targeting the iron sulfur complex assembly (ISCU), which constitutes the core of electron transport complex I–IV ( 56 , 57 ). miR-210 further helps in the stabilization of HIF-1 by suppressing the expression of glyceroldehyde 3-phosphate dehydrogenase like 1 enzyme ( 58 , 59 ). Functional activity of HIF-1 is strengthened by down regulation of FIH by miR-31 ( 59 ). Similarly, a recent study has established that miR-183 helps in stabilizing HIF-1 by targeting isocitrate dehydrogenase 2 (IDH2) ( 60 ). Thus, HIF-1 levels and its function are regulated by a group of microRNAs which fine-tunes hypoxic adaptation ranging from inflammation ( 61 ), angiogenesis ( 62 , 63 ) and tissue remodeling ( 64 , 65 ). Biological consequence of altering miR in diseased tissues has become a reality recently. Locked nucleic acid (LNA) strategy has been developed to bind and neutralize activity of specific miR under clinical settings. LNA-miR-92a ( 66 ) exerted cell-protective, proangiogenic, and anti-inflammatory effects in pig models. Similarly, miR-15 was found to protect against cardiac ischemic injury ( 67 ). These preclinical studies are encouraging and in fact LNA-based miR knock down methods are currently under Phase I/II trials. Clinical success of these methods however depends on effective delivery of LNA-reagents into affected target tissues. Pharmacological stability, tissue penetration, transport across cell membranes, escaping from nucleases and finally specific neutralization of target miR are some of the issues currently being investigated ( 68 – 70 ). Recent advances in the development of nanocarriers and functionalized nanopartlcles have helped in improving microRNA-mediated targeted therapies. These strategies are currently evaluated to knock down miRs (antagomiRs) targeting HIF-1α and VEGF to induce collateral vessel growth in ischemic tissues and to control inflammatory response.

Cells adapt to hypoxia by recalibrating their metabolic needs and activating survival pathways. Availability of oxygen is directly linked to energy homeostasis. Lower oxygen levels compromises the function of mitochondria in generating cellular energy currency, ATP, through oxidative phosphorylation, which is the most efficient way of producing ATP from glucose. Cells activate three major pathways to survive the energy crisis created by hypoxia. The first priority for cells under hypoxic stress is to minimize energy demand by conservation. Low-priority, energy consuming processes such as endocytosis is shut down under hypoxia ( 18 ). Then, energy production by glycolysis is increased. In order to maintain energy balance, alternate sources of energy are also utilized. Finally, hypoxia forces cells to recycle macromolecules (macroautophagy) and organelles such as mitochondria (mitophagy) to make raw materials available for biosynthetic machinery ( 19 , 20 ). Adaptation to nutrients and oxygen availability is an evolutionarily conserved process. From unicellular organisms to mammals, oxygen levels are continuously monitored by oxygen sensing machinery, which reprograms expression of several genes to adapt to changing conditions. Failure to do so will be catastrophic. In this review, we will focus on how mammalian cells respond to hypoxia with a complex regulatory system involving epigenetic and genetic pathways. Semenza and his colleagues discovered hypoxia inducible factor, HIF-1, while studying the regulation of erythropoietin (EPO) gene expression during hypoxic stress. Their pioneering studies identified cis-acting enhancer element in the promoter of EPO ( 21 ). Subsequent investigations revealed that HIF-1 binds to a consensus sequence, NCGTG, termed as hypoxia response element (HRE). HRE sites are present in the promoter of hundreds of genes, which are functionally related to hypoxic adaptation. For example, HIF induces expression of GLUT-1, glucose transporter that increases the intracellular levels of glycolytic substrate. At the same time, many of the enzymes needed for increased breakdown of sugar, hexokinase-1, glucose 6-phosphate isomerase, phophofructokinase, PFK1, the rate limiting enzyme of glycolysis, phosphoglycerate mutase-1 (PGAM1), pyruvate kinase type M-2 (PKM2), lactate dehydrogenase (LDHA) are increased by HIF-1-mediated transactivation. While increasing glycolysis, HIF-1 also blocks mitochondrial function and redirects the utilization of glutamine to citrate by upregulating isocitrate dehydrogenase-1 (IDH1) and aconitase1. This pathway is necessary to maintain fatty acid synthesis during hypoxia since, the production of acetyl-coA from pyruvate is prevented during hypoxia by the inhibition of pyruvate dehydrogenase enzyme by pyruvate dehydrogenase kinase1 (PDK1). Details of metabolic pathways regulated by HIF-1 under hypoxia is elegantly summarized in a recent review by Semenza ( 22 ). Two recent studies have further confirmed the critical role of HIF-1 in glutamine utilization under hypoxia ( 23 , 24 ). Reductive carboxylation is a preferred pathway to generate citrate and acetyl-coA from glutamine metabolism. Generally, cells use glutamate dehydrogenase 1 (GLUD1) to produce a-ketoglutarate. However, some malignant cells use a non-canonical to convert glutamine derived aspartate to oxaloacetate/a-ketoglutarate. This pathway was mediated by aspartate transaminase (GOT1) in a HIF-1 dependent manner. HIF-1 seems to be necessary and sufficient for reductive carboxylation, an important metabolic adaptation for the synthesis of fatty acids under hypoxia. Fatty acid precursors are needed for the synthesis of inflammatory mediators. In addition to maintaining energy homeostasis, HIF is also directly responsible for the increased production of EPO, vascular endothelial growth factor (VEGF), necessary for angiogenesis and nitric oxide synthase generating nitric oxide that is responsible for vasodilation and increased blood flow to ischemic tissues. In addition to metabolic and angiogenic regulation, HIF-1 is also responsible for the expression of inflammatory cytokines such as IL-33 in RA synovial fibroblasts. IL-33 binds to IL1Rl1 (ST2) and induce the production of Th-2 cytokines ( 25 ). IL-33 has dual role, intracellular and extracellular activation. NH-2 terminus has NLS motif and the full length IL-33 localizes to nucleus and regulates transcriptional activity ( 26 ). Proteolysis allows extracellular presence of IL-33 and stimulates IL1RL1 receptors on Th-2 cells ( 27 ). It may function as an ‘alarmin’ in directing immune/inflammatory response to the sites of tissue damage, similar to IL-1α and HMGB1 ( 28 ). IL-33 is highly expressed in the heavy endothelial venule endothelial cells (HEV EC). HEV EC is directly linked to extravasation of lymphocytes from circulation into lymph nodes by diapedesis ( 29 ). IL-33 may have an important role in sustained inflammation associated with diseases such as RA and Crohn’s disease. IL-33 plays a significant role in innate immunity as well ( 30 ). Thus HIF is a master regulator of hypoxic adaptation. The repertoire of genes modulated by HIF is being unraveled only recently.