Hypoxia-inducible fam

iverch

Reviewcells [1,2]. Together, cancer cells and host cells form atumor microenvironment that enables tumor initiationand progression. The dependence of cancer on the hostcells suggests that these host cells could also be targetedfor cancer therapy. Again, because conventional cancertherapy was developed without emphasizing the tumormicroenvironment, a major new focus for cancer therapyis to limit cancer development by targeting this microen-vironment.

and activity of pyruvate kinase muscle isozyme 2 (PKM2),HIF-1a can serve as a master switch for oxygen regulationin cellular metabolism [17].

As illustrated in Figure 1, HIF is a heterodimer com-prising a and b subunits. The heterodimers translocateinto the nucleus, where they interact with specific DNAsequences called HIF-responsive elements (HREs). Bybinding to the HRE, HIF may either activate or repressgene expression. At least three different genes have beenidentified that encode a subunit of HIF, namely HIF1a,HIF2a and HIF3a. All three HIFa subunits heterodimer-ize with a HIF-1b subunit and are subject to post-translational regulations that are dictated by the oxygenconcentration in the environment. Although HIF-3a lacksa transactivation domain and is generally considered to bea negative regulator for HIF-1a and HIF-2a function, anotable exception was reported recently [18]. Despite the

0165-6147/

2015 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tips.2015.03.003

Corresponding author: Liu, Y. (yaliu@cnmc.org).Keywords: leukemia-initiating cells; tumor microenvironment; myeloid-derivedsuppressor cells; dendritic cells; tumor-associated macrophages; cancer immunother-apy.stem cells and inflaGong Peng1 and Yang Liu1,2

1 Institute of Translational Medicine, The First Hospital, Jilin Un2Center for Cancer and Immunology Research, Childrens ReseaUSA

Hypoxia-inducible factors (HIF) mediate metabolicswitches in cells in hypoxic environments, includingthose in both normal and malignant tissues with limitedsupplies of oxygen. Paradoxically, recent studies haveshown that cancer stem cells (CSCs) and activated im-mune effector cells exhibit high HIF activity in normoxicenvironments and that HIF activity is critical in themaintenance of CSCs as well as the differentiation andfunction of inflammatory cells. Given that inflammationand CSCs are two major barriers to effective cancertherapy, targeting HIF may provide a new approach todeveloping such treatments.

Challenges in targeting the tumor microenvironmentAn important paradigm shift in cancer research has beenthe realization that cancer tissue is not a homogenouspopulation of clonally expanded cancer cells [13]. It hasbeen established in multiple cancer types that cancer cellsare hierarchical: while a small subset of CSCs have a highcapacity for self-renewal and are responsible for initiatingcancer, the bulk of cancer cells lack self-renewal andcancer-initiating capacity [3,4]. Perhaps because conven-tional therapeutic approaches were not developed to targetCSCs, many such cells are enriched by conventional cancertherapy [5,6]. The ineffective elimination of CSCs is con-sidered a major cause of cancer relapse following conven-tional therapy and, thus, is a major barrier to effectivecancer treatment [7].

In addition to heterogeneity among transformed can-cer cells, cancer tissues also comprise noncancerous host374 Trends in Pharmacological Sciences, June 2015, Vol. 36, No. 6ctors in cancermation

rsity, Changchun, China Institute, Childrens National Medical Center, Washington, DC,

Based on these considerations, it would be of interest toidentify druggable targets that are critical for CSCs andtumor-promoting microenvironments. In this context,studies have highlighted the selective activation of theHIF pathway and metabolic switch of CSCs [810]. Re-markably, recent studies demonstrated that HIF may havea major role in inflammation, including the adaptive andinnate inflammatory responses [1113]. The shared re-quirement for HIF in CSCs and inflammatory cells raisedthe interesting prospect that CSCs and inflammation, twoimportant challenges in cancer therapy, may be addressedby targeting HIF. Here, we review the critical role for HIFin immunology and cancer biology, with a focus on thepotential cross-fertilization of HIF research on cancer cellsand host inflammatory cells. We also explore the transla-tional potential of this new concept. Readers are referred tooutstanding recent reviews for the general concept andinvolvement of HIF in cancer biology and immunology[11,14,15].

HIF and cancer: an overviewThe modes of energy production in a cell are usuallydictated by the oxygen levels in the environment: oxidativephosphorylation occurs in a well-oxygenated environment(normoxia), while glycolysis is switched on when oxygenlevels drop below 1% (hypoxia) [15]. In a search for thefundamental mechanism of the oxygen-mediated metabol-ic switch, Gregg Semenza and colleagues identified HIF-1aan oxygen-sensitive transcriptional activator [16]. Recentstudies suggest that, by directly regulating the expression

P402OH P564OH

P402 P564

PHD1, 2, 3

VHL

O2 O2

Ub Ub

ProteasomepolyUb

HIF-1 HIF-2HIF-1

HRE

P405 P531

P405OH P531OH

O2

TRENDS in Pharmacological Sciences

Figure 1. Regulation of hypoxia-inducible factors (HIF) activity in response to oxygen

levels. Under hypoxic conditions, stabilized HIF-1a and HIF-2a dimerize with HIF-1b.

The heterodimers translocate into the nucleus to regulate gene transcription. In the

presence of oxygen, HIF-1a and HIF-2a proteins are hydroxylated by prolyl

hydroxylase domain proteins (PHD) 13 at the proline residues indicated.

Hydroxylated HIF is recognized by von Hippel-Lindau tumor suppressor, E3

ubiquitin protein ligase (VHL), which causes polyubiquitinylation (polyUb) and

Reviewgeneral similarity in regulation and function between HIF-1a and HIF-2a, they differ in their sensitivity to oxygendeprivation, and target gene binding and tissue distribu-tion [14].

The critical role of HIF in cancer biology was establishedwhen a major renal tumor suppressor gene von Hippel-Lindau tumor suppressor, E3 ubiquitin protein ligase(VHL) was identified as the E3 ligase responsible ubiqui-tinylation of HIF-1a and HIF-2a [1923]. VHL recognizeshydroxylated residues at Pro402 and/or Pro562 in HIF-1aand Pro 405 and Pro 531 in HIF-2a, by means of the prolylhydroxylase domain protein (PHD) [24]. VHL is inacti-vated in most renal cancer samples, leading to increasedexpression of the HIF-1a and HIF-2a proteins [1923]. Mutations of PHD protein-encoding genes, EGLN1(PHD2), EGLN2 (PHD1), and EGLN3 (PHD3) have beenobserved in various cancers at low frequency (http://www.cbioportal.org/). The function of PHD is enhanced by iso-citrate dehydrogenase 1 (IDH1) and/or IDH2 [25,26], whichare mutated at a combined frequency of approximately15% in patients with acute myeloid leukemia (AML) [2730] and low-grade glioma or glioblastoma [31]. Two reportssuggested that IDH mutations lead to increase in HIF-1aaccumulation [25,26], while a more recent study suggestsotherwise [32].

Overexpression of HIF-1a and/or HIF-2a is a marker forpoor prognosis in most cancer types tested, includingcommon cancers, such as breast, prostate, colon, hepato-cellular, pancreatic, brain, and ovarian cancers, as well asa host of less common cancers [14]. In transgenic mouse

proteasome-mediated degradation of HIF. Abbreviation: HRE, HIF-responsive

element.cancer models, heterozygous deletion of Hif1a reduced thegrowth of thymic lymphoma [33], while short hairpin(sh)RNA silencing of Hif1a and overexpression of VHLstrongly reduced leukemia-initiating activity [9].

Despite the overwhelming association between HIFlevels in cancer tissue and poor prognosis of patients withcancer, conflicting reports have emerged in renal cancer,where overexpression of HIF-1a has been associated witheither a poor or favorable prognosis, depending on themethods used to evaluate HIF-1a levels [34,35]. In mousecancer models, recent studies suggest that, while localdeletion of Hif1a in lung tissue had no impact on KRas-driven lung cancer, deletion of Hif2a paradoxically accel-erated lung cancer development [36]. More recently, broaddeletion of Hif1a in adult mice, including in the cells thatgive rise to leukemia, promoted, rather than suppressedMll-AF9a-induced leukemia [37]. These contradictingobservations may be explained by the broad spectrum ofHIF target genes: the cancer-promoting effect is exempli-fied by HIF-mediated induction and function of PKM2,vascular endothelial cell growth factor (VEGF), and multi-drug resistance gene (MDR1), while its function as a tumorsuppressor can be explained by both its transcriptional andnontranscriptional activity.

PKM2 and metabolic switches in cancer cellsPKM2 is the final enzyme in glycolysis and comprises twoisoforms that arise from alternative splicing. Most tissuesexpress the PKM1 isoform with products directed intooxidative phosphorylation to efficiently produce ATP. Bycontrast, PKM2 may exist either in tetrameric or dimericforms to direct oxidative phosphorylation or glycolysis,respectively. Given that PKM2 exists predominantly inits dimeric form in cancer cells, it contributes to a highrate of aerobic glycolysis in cancer cells regardless ofhypoxia, a phenomenon known as the Warburg effect.HIF regulates PKM2 through two mechanisms [17]: first,HIF-1a can stimulate expression of PKM2. Second,PKM2 and HIF-1a form heterodimers and migrate intothe nuclei, where they act as a master switch to over-express genes crucial for a robust glycolysis that producesboth energy and metabolic intermediates for biosynthe-sis. Therefore, HIF-1a has a critical role in metabolicswitches in cancer cells. In addition to PKM2, HIF-1a hasbeen shown to antagonize cMyc activity by inducingexpression of MXl1, thus reducing mitochondrial biogen-esis [38].

VEGF and cancer neoangiogenesisThe increase in tumor volume demands a correspondingincrease in angiogenesis. By regulating VEGF expression,both cellular and viral oncogenes not only regulate thegrowth of cancer cells, but also allow cancer progression inthe host. Optimal VEGF expression was found to dependon both hypoxia and oncogenes [39]. These observationsled to the identification of HIF-1a as a major regulatorof VEGF expression [40,41]. The HIF-1ap300/CREB-binding protein (CBP) complex binds to a HIF-responsive

0

Trends in Pharmacological Sciences June 2015, Vol. 36, No. 6element in the 5 promoter region of the gene encodingVEGF [40,41]. This HRE sequence was specifically tar-geted by echinomycin [9,42].

375

MDR1The link between HIF and MDR1 was first reported innontransformed endothelial and epithelial cell lines[43]. The authors showed that HIF1a was responsiblefor hypoxia-induced MDR1 expression through stimula-tion of the MDR1 promoter sequence. Subsequently, it wasshown in head and neck cancer cell lines that inhibition ofHIF-1a increased cancer cell sensitivity to paclitaxel[44]. The contribution of HIF to multidrug resistance isnow substantiated in many types of cancer, includinggastric cancer, colon cancer, multiple myeloma, AML,laryngeal cancer, and sacral chordoma [4551]. Theseobservations explain the poor prognosis of patients withcancer overexpressing HIF and suggest an important rolefor HIF inhibitors in combinatorial cancer therapy.

Tumor growth inhibition by HIF-1aDespite compelling evidence for the oncogenic effect of HIFfamily members, overactivation of HIF may cause growtharrest and/or apoptosis. At least three mechanisms havebeen proposed to explain the tumor suppressor activity ofHIF-1a. First, HIF-1a has been shown to stabilize p53,perhaps by binding to Mouse double minute 2 homolog(MDM2) [52]. Second, among HIF-1a targets are genescapable of inducing apoptosis, including RPT801 [53],NIP3 [54,55], and NIX [55]. Third, HIF-1a has been shownto antagonize Myc by preventing its repression of p21[56]. In combination, these mechanisms may reconcilethe inconsistencies in the mouse genetic data that suggestthat Hif1a either promotes or suppresses tumor growthdepending on the tumor types investigated [33,37].

HIF and CSCsAn important advance in cancer biology was made with thedemonstration that somatically transformed cancer cellsare heterogeneous in establishing cancer in a new host. Inmany solid and hematological cancers, a small subset ofcancer-initiating cells can be prospectively isolated basedon either cell surface phenotypes or their ability to excludefluorescent dyes (side population). Although it has beensuggested that therapeutic elimination of CSCs holds thekey to reducing cancer relapse and drug resistance, fewdruggable targets have been identified that provide ameans for the selective elimination of CSCs [5760]. Asreviewed below, accumulating data suggest that HIF willemerge as a long-sought target.

Glioma stem cells and HIFThe involvement of the HIF pathway in glioma stem cellswas first reported by Li et al. [8]. Using xenograft glioma-initiating and in vitro tumorosphere formation assays andCD133 as the CSC markers, these authors observed sig-nificant enhancement of stem cell activity when the stemcells were cultured under a hypoxic environment. Inter-estingly, stem cell activity under both normoxic and hyp-oxic environments is reduced when either HIF1a or HIF2aare silenced by shRNA. Given that HIF2a mRNA levelscorrelate with glioma activity, progression, and prognosis,

Reviewthe authors emphasized that HIF2a is critical for gliomastem cell activity. However, shRNA silencing showed anequally important role for HIF1a in CSC activity. The lack

376of correlation between HIF1a mRNA levels and stem cellactivity may simply reflect the fact that HIF-1a proteinlevels are regulated by post-transcriptional mechanisms.The HIF targets responsible for CSC activity remain to beidentified.

Acute lymphocytic leukemiaEarly work on CSCs was criticized for using xenografttransplantation because the CSC activity (i.e., tumor-initiating activity) assayed in the xenogeneic host maybe artificially affected by rejection of human cells by theinnate immunity of the recipient mice [61]. Wang andcolleagues used a syngeneic transplantation model to dem-onstrate that the c-Kit+Sca-1+ population in their mousemodel of acute lymphocytic leukemia (ALL) were leuke-mia-initiating cells [9]. Given that c-Kit+Sca-1+ cells are anecessary and sufficient population for an in vitro colony-forming unit (CFU), the authors used this model to identifyinhibitors that may target ALL stem cells. They found thatechinomycin, a natural product that binds to HRE and,thus, inhibits HIF activity, efficiently eliminated CFU withan IC50 of 30100 pM. In vivo, low-dose administration ofechinomycin (10 mg/kg or 30 mg/m2) cured 100% of synge-neic mice that had received lethal doses of leukemia cells.The involvement of HIF-1a is substantiated by three linesof evidence. First, Hif1a, but not Hif2a, is expressed atsubstantially higher levels in the leukemia stem cell (LSC)population compared with the bulk of leukemia blasts.Second, shRNA silencing of Hif1a reduced CFU and leu-kemia-initiating activity. Third, Hif-1a was found to en-hance Notch signaling by preventing negative feedbackregulation of the Hes family bHLH transcription factor 1(Hes1) gene, a key Notch target known to be involved instem cell self-renewal. However, it remains to be estab-lished whether upregulation of Hes1 is responsible for HIF-mediated maintenance of LSC.

Using a HIF reporter, Wang et al. showed that, undernormoxic conditions, HIF is active exclusively within thec-Kit+Sca-1+ LSC subset [9]. This high HIF activity wasdue to the selective loss of VHL expression in the stem cellsubset, which is critical for LSC function, given that ectopicexpression of Vhl eliminates LSC [9].

AMLHistorically, the prospective identification of AML stemcells by John Dick and colleagues in 1994 marked therevival of the CSC concept [4]. Several groups have shownthat the abundance of AML stem cells is a biomarker forpoor prognosis [6264]. Gene expression signatures de-rived from the comparison of AML stem cells and AMLblasts have proven valuable in predicting the outcome ofnewly diagnosed patients [65]. An important prediction ofthe CSC concept is that therapeutics that selectively elim-inate stem cells should prevent drug resistance and cancerrelapse, the two most pressing issues in cancer therapy.Given that echinomycin can selectively eliminate AMLstem cells [9], the planned clinical trials using echinomycin[66] may provide an opportunity to test this concept.

Trends in Pharmacological Sciences June 2015, Vol. 36, No. 6Using the AML stem cell markers identified by Dicksgroup, Wang and colleagues [9] showed that HIF1a mRNAand protein are overexpressed in human AML stem cells

compared with the bulk of AML blasts. In both nave andtreated AML samples, echinomycin efficiently inhibitedCFU activity, with an IC50 of approximately 100 pM. Moreimportantly, echinomycin was 1001000-fold more effi-cient in inducing apoptosis of CD34+CD8 AML stem cellscompared with the bulk of AML blasts. To test the thera-peutic potential, the authors established human AML inNOD.SCID mice according to methods described by theDick laboratory [4] and treated the mice with a low dose ofechinomycin [9]. Short-term treatment with echinomycinnot only reduced leukemia blast burden, but also prefer-entially reduced the frequency of AML stem cells withinthe leukemia cells, marking a major difference comparedwith conventional therapeutics. As evidence of functionalinactivating AML stem cells, the remaining AML blast(s)could no longer initiate AML in the new host.

To test whether echinomycin can be used to treat re-lapsed AML, Wang and coworkers [67] used the relapsedAML from mice with heterozygous knock-ins of two genesthat are frequently mutated in human AML: FLT3ITD andMLLPTD (MllPTD/WT:Flt3ITD/WT). The MllPTD/WT:Flt3ITD/WT

mice developed spontaneous AML and responded to treat-ment with a DNA hypomethylating agent and/or a histonedeacetylase (HDAC) inhibitor; however, the AML invari-ably relapsed [68]. The authors transplanted the relapsedAML from CD45.2 mice into CD45.1 mice and followed theexpansion and therapeutic response using CD45.2 as aleukemia marker [64]. They observed that short-termtreatment with echinomycin cured 4060% of mice witha high burden of relapsed AML at the time of treatment.Again, the bone marrow from the cured mice did notharbor dormant AML stem cells, as evidenced by theirinability to initiate AML in the new hosts. Surprisingly,echinomycin-treated bone marrow cells are fully compe-tent in hematopoiesis in transplantation. Therefore, ther-apeutic elimination of CSCs can be achieved in relapsedAML without significant adverse effects on tissue stemcells. Thus, although CSCs and tissue stem cells may shareself-renewal programs, the therapeutic elimination of CSCscan be achieved without harming tissue homeostasis.

Chronic myeloid leukemiaChronic myeloid leukemia (CML) is a clonal myeloprolif-erative disorder that results from an acquired geneticchange in a single hemopoietic stem cell. The hallmarkof CML is the generation of the BCR-ABL chimera protein,a constitutively active tyrosine kinase, as a result of genetranslocation [69]. Induction of BCR-ABL induces expres-sion of HIF-1a and its target gene VEGF [70]. Kinaseinhibitors (such as imatinib) fail to eradicate CML LSCs.Instead, the targeted therapy selected for imatinib-resistant cells with high levels of BCR-ABL and activeHIF-1a, resulting in anaerobic glycolysis and increasedsurvival in vitro [71]. Using a mouse model of CML, Zhanget al. showed that Hif1a-deficient HSPCs expressingBCR-ABL failed to generate CML in secondary recipients.Deletion of Hif1a prevents the propagation of CML byimpairing cell cycle progression and inducing apoptosis

Reviewof LSCs [72]. In addition, recent studies also showed thatHIF-1a supports the maintenance of CML LSCs despiteeffective BCR-ABL1 inhibition in hypoxic environments[73]. These observations indicated that Hif-1a is crucial forthe maintenance of CML LSCs.

Taken together, the above examples illustrate that HIFhas a critical role in stem cells for both solid tumors andhematological malignancies, serving as a druggable targetfor the therapeutic elimination of CSCs. The fact that HIFis active in CSCs regardless of hypoxia suggests that thesecells have acquired mechanisms to protect HIF undernormoxic conditions.

Breast CSCsBreast cancer is the first solid tumor in which CSCs wereprospectively isolated [3]. Using a xenograft model ofhuman breast cancer cell lines, Conley et al. observed thatantiangiogenic agents increased the proportion of breastCSCs (BCSCs) by inducing hypoxia. The impact of hypoxiais mediated by HIF-1a, but not HIF-2a, and correlates withWnt signaling in BCSC [74]. Using the mouse mammarytumor virus polyoma virus middle T (MMTV-PyMT)oncogene-induced breast cancer model, Schwab et al. [75]demonstrated that conditional deletion of Hif1a in mamma-ry epithelial cells causes reduced tumor growth and metas-tasis. Interestingly, this depressed tumor growth andmetastasis corresponded to a reduction in the number andfunction of BCSC [75].

A key issue is how HIF-1a induces BCSC. One keyregulator of BCSC activity is a Hippo pathway effectormolecule called TAZ [76]. By comparing gene expressionprofiles among1600 cases of human breast cancer samples,Xiang et al. [77] showed a significant correlation betweenknown HIF target genes and TAZ, raising the intriguingpossibility that TAZ is a direct target of HIF-1a. Thishypothesis is validated by shRNA silencing, chromatin-immunoprecipitation, promoter activity, TAZ target geneco-expression, and hypoxia responses. In addition to con-trolling TAZ expression, HIF-1a also activates transcrip-tion of the Siah E3 ubiquitin protein ligase 1 (SIAH1) gene,which encodes a ubiquitin ligase that is required forproteasome-dependent degradation of large tumor sup-pressor kinase 2 (LATS2). Degradation of LATS2 allowednuclear localization of TAZ. Apart from the expression andfunction of TAZ, hypoxia also regulates the interactionbetween BCSC and mesenchymal stem cells (MSC) andtumor-associated macrophages (TAM) through two feed-forward mechanisms [78]. First, HIF-1a induces expres-sion of chemokine (C-X-C motif) ligand 16 (CXCL16) torecruit MSC. Second, MSCs produce chemokine (C-C mo-tif) ligand 5 (CCL5) to recruit TAM. Therefore, BCSC-intrinsic signaling not only maintains the number andfunction of the BCSC, but also drives formation of thetumor microenvironment.

HIF in adaptive and innate inflammationHypoxia-resistant HIF activity is not limited to cancercells. Accumulating evidence shows that HIFs have a rolein the regulation of innate and adaptive immune cells thatnormally reside in normoxic environments. ActivatedT cells express two isoforms of HIF-1a: a short form called

Trends in Pharmacological Sciences June 2015, Vol. 36, No. 61.1 and a longer form called 1.2 [79]. The shorter 1.1 formappears to suppress the production of inflammatory cyto-kines by T cells [80]. Accumulating data demonstrate that

377

HIF is a key regulator of regulatory T cell (Treg)/T helper(TH)-17 development, the differentiation and function ofcytotoxic T lymphocytes, (CTL), and innate immune effec-tors, such as dendritic cells, neutrophils, and myeloid-derived suppressor cells (MDSCs).

Treg/TH17 cell differentiationAlthough Th17 and Treg cells share important pathways intheir differentiation from nave T cells, they eventuallybifurcate into distinct phenotypes with opposite activities,with TH17 cells being pro-inflammatory and Tregs beinganti-inflammatory. Recent studies suggest that the func-tional switch between these cell types is mediated by HIF-1a.

HIF-1a activates Retinoid-Acid Receptor-related Or-phan Receptor (ROR)-gt transcription and forms a tertiarycomplex with RORgt and p300 to activate the IL-17A gene.By contrast, Hif-1a binds to forkhead box P3 (Foxp3) andcauses its degradation, resulting in a blockade of Tregdifferentiation. As a result of altered Th17/Treg differenti-ation, mice with HIF-1a-deficient Treg are resistant toexperimental autoimmune encephalomyelitis [81]. Whennave T cells were cultured in the presence of transforminggrowth factor (TGF) b and IL6, they showed upregulationof the glycolytic activity and induction of glycolyticenzymes. A critical role for glycolysis in Th17 developmentis also suggested, given that blocking glycolysis with2-dexoglucose inhibited TH17 cell development. Consis-tent with a critical role for HIF in Th17 differentiation,HIF-1a was selectively expressed in TH17 cells, whileHif1a deficiency inhibited TH17 differentiation [82]. Thefunction of HIF-1a in TH17 was not restricted to regulationof the metabolic switch. HIF-1a also controlled the expres-sion of antiapoptotic Bcl-2 family genes in collaborationwith Notch signaling and, thus, supports TH17 survival inhuman samples [83]. Taken together, these data highlightHIF-1a-dependent pathways in regulating the balancebetween the differentiation of TH17 and Treg cells.

Despite the in vitro effect of HIF-1a on Treg differenti-ation, the inhibitory function of HIF on Treg function in thetumor microenvironment remains to be substantiated.Surprisingly, HIF promotes differentiation of CD4+/CD25+Treg cells in the case of hypoxia caused by rapidtumor progression [84]. Tregs may be enriched in cancertissue by HIF-1a-dependent production of chemokines,such as CCL28 [85]. Additional studies are needed todetermine whether the apparent contradiction was attrib-utable to the tumor microenvironment, and if so, how thetumor microenvironment alters the role of HIF-1a in Tregdifferentiation.

CTL differentiation and functionCTLs are key adaptive effectors in the elimination ofintracellular pathogens and tumor cells. HIF-1a deter-mines CTL fate by regulating transcription of genes encod-ing glucose transporters, rate-limiting glycolytic enzymes,cytolytic effector molecules, and essential chemokine andadhesion receptors that regulate T cell trafficking [86]. Ac-

Reviewtivated T cells express high levels of HIF-1a proteins evenunder normoxia, although hypoxia further elevates HIF-1alevels [86,87]. The induction of HIF-1a is controlled by

378mammalian target of rapamycin complex 1 (mTORC1)[86]. Genetic studies in mice demonstrated that deletionof Vhl in CTLs inhibited persistent viral infection andneoplastic growth, while that of Hif1a attenuated CTLeffector function during viral infection [87]. Surprisingly,deletion of Hif1a in activated T cells also enhanced pro-duction of inflammatory cytokines. These data argue for anegative regulatory role for HIF-1a in T cell-mediatedinflammatory responses [80,88]. Further studies are need-ed to shed light on T cell-intrinsic HIF-1a function in thetumor microenvironment.

Dendritic cell activationAs the primary antigen-presenting cells, dendritic cells(DCs) have a crucial role in linking the innate and theadaptive immune systems. LPS and hypoxia lead to in-creased glycolytic activity and increased DC maturation.The role for HIF-1a has been confirmed, given that knock-ing down of HIF-1a in DCs significantly reduced glucoseutility and inhibited DC maturation, as evidenced byreduced stimulation of allogeneic T cells [89]. While a rolefor HIF-1a in tumor-infiltrating DC has not been investi-gated, the induction of B7 homolog 1 (B7H1/PD-L1) byHIF-1a [90,91] and immune suppression by B7H1-expres-sing myeloid DC in human cancer [92] suggest that HIF-1aexpressed in tumor-infiltrating DC contributes to the im-mune-suppressive tumor microenvironment.

Tumor-associated myeloid suppressor cellsTumor growth alters myeloid development in the host,with significant phenotypic and functional changes inthe myeloid compartment [93]. A major change in thetumor-bearing host and the tumor microenvironment isthe expansion of myeloid suppressor cells, including tu-mor-associated macrophages (TAM), MDSC, and myeloiddendritic cells (mDC) [93]. Accumulating data demonstratethat HIF-1a contributes to the expansion, differentiation,and effector function of myeloid suppressor cells within thetumor microenvironment.

As discussed above, HIF-1a may contribute to mDC-mediated immune suppression by regulating the expres-sion of B7H1/PD-L1. Both TAM and MDSC suppress tumorimmunity by producing soluble factors such as nitric oxide(NO) and ROS [93,94], and through expression of cellsurface proteins, such as B7H1/PD-L1. HIF-1a, but notHIF-2a, upregulates the expression of B7H1/PD-L1 bybinding to the HRE of the PD-L1 proximal promoter toenhance expression of PD-L1 [90]. In addition, HIF-1a isnecessary for the optimal expression of arginase and in-ducible nitric oxide synthase (iNOs) in both TAM andMDSC and, thus, is responsible for NO production [95]. In-terestingly, MDSC have been shown to differentiate intoTAM in the tumor microenvironment, and this process isrequires expression of HIF-1a MDSC [95].

The contribution of HIF to the tumor microenvironmenthas been demonstrated mostly using transplantable tumormodels. As a notable exception, myeloid-specific HIF-1adeletion in a transgenic model of breast cancer reduced

Trends in Pharmacological Sciences June 2015, Vol. 36, No. 6tumor growth and progression [96]. This was achievedwithout alternation of VEGF-A levels and vascularization.To test whether this deletion affects macrophage functions,

the authors compared wild type and Hif1a/ bone mar-row-derived macrophages under normoxic and hypoxicculture conditions. The authors observed that hypoxiaexacerbated inhibition of T cell proliferation and that suchexacerbation was Hif1a dependent. Given that TAM areknown to inhibit T cell function [93], it would be of interestto determine whether the delay in tumor growth wasattributable to Hif1a deletion in TAM.

Epigenetic and genetic mechanisms for HIF activitiesunder normoxia: cross-fertilization between cancerbiology and immunologyThe critical role for HIF in cancer cell metabolism, CSCfunction, and immunity raised a fundamental but largelyunanswered question: how is HIF activity maintained incirculating LSCs, cancer cells, and hematopoietic cells in anormoxic environment? Accumulating data indicate thatHIF activation is stimulated by increased transcriptionand translation of HIF1a and inactivation of oxygen-mediated HIF degradation, through both genetic and epi-genetic mechanisms.

Genetic mechanisms taught by cancer: the AMLexample

ReviewThe accumulation of HIF1a protein and its transcriptionalactivity are strictly regulated. As shown in Figure 2, manyof these mechanisms are targeted in AML, and severalgenomic alterations in AML have either been shown, orhave the potential, to enhance HIF activity. First, AML-associated mutations seem to target the canonical HIFdegradation pathway. Under normoxia, HIF-1a is degrad-ed by VHL, which recognizes HIF-1a when it is hydroxyl-ated at Pro402 and/or Pro562 by PHD [24]. Mutations ofgene encoding PHD have been observed in AML only at low

UbHIF1

Degradaon

IHD

VHL

mIHD

2HGmTOR

PI3K/AKT

O2+

K

G

PhDHIF1

Hydroxylated HIF1p53 MDM2

FLT3ITD/TKD

NPM/p14ARF

10%

30% NC AML

2535%

1015%

H d

HI

F1

mIH

Ub

2

d

ITD

TRENDS in Pharmacological Sciences

Figure 2. Prevalent mutations in acute myeloid leukemia (AML) potentially affect

hypoxia-inducible factor (HIF) levels in cancer cells. Mutations of genes encoding

isocitrate dehydrogenase (IDH) 1 and 2, which are mutated in 1015% of AML

cases, inactivate prolyl hydroxylase domain protein (PHD) to suppress HIF

hydroxylation. Fms-related tyrosine kinase 3 (FLT3) mutations (found in 2535%

of AML cases) stimulates the phosphatidylinositol 3-kinase (PI3K)/AKT pathway to

enhance HIF translation. Nucleophosmin (NPM) mutation, which was found in 30%

of AML with normal karyotype (NC), reduces p14ARF, and HIF inhibitors. TP53mutation reduces expression of Mouse double minute 2 homolog (MDM2), an

alternative E3 ubiquitinylation (Ub) ligase E3 for HIF. This in turn may reduce HIF

ubiquitinylation.frequency. However, the function of PHD is enhanced byIDH1 and/or IDH2 [25,26], which are mutated at a com-bined frequency of approximately 15% in patients withAML [2730]. One study showed that IDH mutationsincrease HIF1a accumulation [25,26], while a later studysuggested an opposite effect [32]. The differences betweenthe two remain to be reconciled. Second, fms-related tyro-sine kinase 3 (FLT3)-internal tandem repeats (ITD) andtyrosine kinase domain (TKD) mutations activate FLT3[97,98]. Given that FLT3 signaling activates the phospha-tidylinositol 3-kinase (PI3K)mTOR pathway [99] and thatmTOR activation increases HIF1a protein accumulation[100103], it would be intriguing to test whether FLT3mutations cause HIF activation in AML. Third, loss ofTP53 function in tumor cells has been shown to causedefective MDM2-mediated degradation of HIF-1a [104].In addition, because p53 mutations are observed in ap-proximately 10% of AML samples, and this mutation isassociated with a devastating prognosis [105], it would beof interest to determine whether this mutation contributesto increased HIF-1a accumulation in AML. Given thatMDM2 is an E3 ligase for HIF-1a [106,107], p53 mutationsmay increase HIF-1a levels by decreasing MDM2. Fourth,Nucleophosmin 1 (NPM1) is mutated in 3040% of AMLsamples that have a normal karyotype [108,109]. In addi-tion, since NPM1 sequesters and protects p14ARF againstdegradation [110], NPM mutations may stimulate HIFactivities by inactivating the p14ARF-mediated aberrantdelocalization of HIF1a [111].

Epigenetic mechanism of HIF1a activation: unansweredquestions from CSCs and immune effector cellsAs the offspring of CSCs, cancer cells should have inheritedgenetic alterations that enable oxygen-resistant HIF ac-tivity. Therefore, for cancers that show selective activationof HIF within the CSC compartment, including AML, ALL,CLL, and glioma, genetic alterations in HIF regulators donot provide a full explanation. There are at least twomechanisms for selective HIF-1a activation in mouse LSCs[9]: increased Hif1a mRNA accumulation and lack of ex-pression of Vhl. Both Hif1a upregulation and Vhl down-regulation are necessary for the maintenance of LSC.Given that LSCs are a self-renewing population that alsodifferentiates into cancer cells, it is of interest to deter-mine the epigenetic mechanism that ensures the heritablegene expression pattern of Hif1a and Vhl and how thispattern is lost when the LSCs differentiate into cancercells. Elucidation of this mechanism will also help toformally demonstrate the existence of epigenetic programsthat maintain LSC activity by selective activation of HIF-1a in this compartment.

Similarly, how oxygen-resistant HIF-1a activity is in-duced in the immune effector cells remains largely unex-plained. Accumulating data suggest that activation ofeither T cell receptor [81,82,112] or Toll-like receptors(TLR) [113] induce accumulation of Hif1a mRNA, perhapsthrough suppression of miR-200 [112]. However, addition-al mechanisms are needed to explain oxygen-resistant

Trends in Pharmacological Sciences June 2015, Vol. 36, No. 6HIF-1a activity. The AKT-PI3K pathway induces accumu-lation of HIF-1a through activation of mTOR to increasetranslation of HIF-1a [114]. In addition, mTOR activation

379

Review Trends in Pharmacological Sciences June 2015, Vol. 36, No. 6HIF

CTLTregTh17

DC

MDSC

TAM

CSC

CaC

EC

HIF

TRENDS in Pharmacological Sciences

Figure 3. A central role for hypoxia-inducible factor (HIF) in the tumor

microenvironment. In addition to its critical role for maintenance and proliferation

of cancer stem cells (CSC) and cancer cells (CaC), HIF promotes the differentiation

and/or proliferation of host cells within the tumor microenvironment, including

endothelial cells (EC), dendritic cells (DC), tumor-associated macrophages (TAM),

and myeloid-derived suppressor cells (MDSC). Apart from regulating cells that affect

T cell function, HIF also directly regulates the activation and differentiation of various

functional subsets of T cells, including cytolytic T lymphocytes (CTL), regulatory T

cells (Treg), and interleukin (IL)-17-producing T cells (Th17).prevents oxygen-induced HIF-1a degradation [102]. In thelatter case, mTOR-induced HIF protection against degra-dation requires functional interaction between mTOR andthe HIF-1a domain that is also involved in hydroxylation ofHIF-1a. Therefore, mTOR may inhibit HIF degradationthrough an as yet undefined mechanism.

Recent studies demonstrated that, similar to adaptiveimmunity, innate effector cells may be also capable ofimmune memory, that is, they can be trained to mounta more robust or much reduced recall response [12,13]. Forinstance, stimulation of monocytes with glucan trained themonocytes to mount a more robust innate immune re-sponse to secondary stimulation in vitro and increasedhost resistance to infections by Candida albicans andStaphylococcus aureus [12]. This training is associatedwith extensive epigenetic alterations in the monocytes[13]. Surprisingly, the training is also associated with ametabolic switch from oxidative phosphorylation to glycol-ysis [12]. Given the critical role of HIF-1a in the metabolicswitch, researchers used a genetic model to determinethe involvement of HIF-1a in the recall response of innateeffectors [12]. Indeed, targeted mutation of Hif1a inmyeloid cells was sufficient to ablate trained immunity.The critical role for HIF-1a in trained immunity and theextensive epigenetic reprogramming during the processraised an interesting issue as to whether epigenetic mech-anisms may be directly responsible for oxygen-resistantHIF activation in monocytes.

Taken together, in both immune effectors and CSCs, asteady state of oxygen-resistant HIF-1a accumulation hasbeen achieved. A major challenge remains in defining how

380this crucial state is achieved through nongenetic, perhapsepigenetic, mechanisms.

Concluding remarksHIF was discovered in the process of investigating howliving organisms adjust to a varying oxygen environment[15]. More recent studies have demonstrated that thispathway is operative even under normoxic environmentand that such oxygen-resistant function is critical in cancerbiology and immunology. HIF activation is responsible fora metabolic switch that allows more sustained prolifera-tion with less DNA damage. These features are importantfor the maintenance of CSCs. In the immune system,activation and/or function reprogramming of immuneeffectors is often imprinted through oxygen-resistantHIF accumulation. The converging feature suggests thatcross-fertilization between cancer biologists and immunol-ogists may bring new insights into the mechanism of HIFregulation and new approaches for cancer therapy.

As summarized in Figure 3, cancer cells and immuneeffectors are major components of the tumor microenviron-ment. In this environment, the immune system not onlyfails to defend the host against cancer, but also oftenactively aids carcinogenesis, with functions ranging frompromoting CSC activity [115,116] and angiogenesis [117],to facilitating malignant transformation [118] and metas-tasis [119]. Therefore, effective control of the tumor micro-environment will likely complement the traditionalapproach of cancer therapy, aiming at eliminating cancercells. Given that most contributors to the tumor-promotingmicroenvironment rely on HIF, it is likely that targetingHIF will not only aid in the therapeutic elimination ofCSCs, but also result in depriving cancer cells of a habit-able microenvironment.

AcknowledgmentsWe thank Christopher Bailey for editorial assistance. This work issupported by Jilin University First Hospital and grants from NationalCancer Institute, USA (CA183030 and CA171972 to Y.L.).

References1 Hanahan, D. and Weinberg, R.A. (2000) The hallmarks of cancer. Cell100, 5770

2 Hanahan, D. and Weinberg, R.A. (2011) Hallmarks of cancer: the nextgeneration. Cell 144, 646674

3 Al-Hajj, M. et al. (2003) Prospective identification of tumorigenicbreast cancer cells. Proc. Natl. Acad. Sci. U.S.A. 100, 39833988

4 Lapidot, T. et al. (1994) A cell initiating human acutemyeloid leukaemia after transplantation into SCID mice. Nature367, 645648

5 Bao, S. et al. (2006) Glioma stem cells promote radioresistanceby preferential activation of the DNA damage response. Nature444, 756760

6 Terpstra, W. et al. (1996) Fluorouracil selectively spares acute myeloidleukemia cells with long-term growth abilities in immunodeficientmice and in culture. Blood 88, 19441950

7 Wicha, M.S. et al. (2006) Cancer stem cells: an old ideaa paradigmshift. Cancer Res. 66, 18831890 discussion 18951886

8 Li, Z. et al. (2009) Hypoxia-inducible factors regulate tumorigeniccapacity of glioma stem cells. Cancer Cell 15, 501513

9 Wang, Y. et al. (2011) Targeting HIF1alpha eliminates cancer stemcells in hematological malignancies. Cell Stem Cell 8, 39941110 Wang, Y.H. et al. (2014) Cell-state-specific metabolic dependency inhematopoiesis and leukemogenesis. Cell 158, 13091323

Review Trends in Pharmacological Sciences June 2015, Vol. 36, No. 611 Palazon, A. et al. (2014) HIF transcription factors, inflammation, andimmunity. Immunity 41, 518528

12 Cheng, S.C. et al. (2014) mTOR- and HIF-1alphamediated aerobicglycolysis as metabolic basis for trained immunity. Science 345,1250684

13 Saeed, S. et al. (2014) Epigenetic programming of monocyte-tomacrophage differentiation and trained innate immunity. Science345, 1251086

14 Keith, B. et al. (2012) HIF1alpha and HIF2alpha: sibling rivalry inhypoxic tumour growth and progression. Nat. Rev. Cancer 12, 922

15 Semenza, G.L. (2012) Hypoxia-inducible factors in physiology andmedicine. Cell 148, 399408

16 Wang, G.L. et al. (1995) Hypoxia-inducible factor 1 is a basic-helix-loop-helix-PAS heterodimer regulated by cellular O2 tension. Proc.Natl. Acad. Sci. U.S.A. 92, 55105514

17 Luo, W. et al. (2011) Pyruvate kinase M2 is a PHD3-stimulatedcoactivator for hypoxia-inducible factor 1. Cell 145, 732744

18 Zhang, P. et al. (2014) Hypoxia-inducible factor 3 is an oxygen-dependent transcription activator and regulates a distincttranscriptional response to hypoxia. Cell Rep. 6, 11101121

19 Kondo, K. et al. (2002) Inhibition of HIF is necessary fortumor suppression by the von Hippel-Lindau protein. Cancer Cell 1,237246

20 Latif, F. et al. (1993) Identification of the von Hippel-Lindau diseasetumor suppressor gene. Science 260, 13171320

21 Maranchie, J.K. et al. (2002) The contribution of VHL substratebinding and HIF1-alpha to the phenotype of VHL loss in renal cellcarcinoma. Cancer Cell 1, 247255

22 Pereira, T. et al. (2003) Identification of residues critical for regulationof protein stability and the transactivation function of the hypoxia-inducible factor-1alpha by the von Hippel-Lindau tumor suppressorgene product. J. Biol. Chem. 278, 68166823

23 Zbar, B. (1995) Von Hippel-Lindau disease and sporadic renal cellcarcinoma. Cancer Surv. 25, 219232

24 Semenza, G.L. (2010) HIF-1: upstream and downstream of cancermetabolism. Curr. Opin. Genet. Dev. 20, 5156

25 Xu, W. et al. (2011) Oncometabolite 2-hydroxyglutarate is acompetitive inhibitor of alpha-ketoglutarate-dependent dioxygenases.Cancer Cell 19, 1730

26 Zhao, S. et al. (2009) Glioma-derived mutations in IDH1 dominantlyinhibit IDH1 catalytic activity and induce HIF-1alpha. Science 324,261265

27 Ward, P.S. et al. (2010) The common feature of leukemia-associatedIDH1 and IDH2 mutations is a neomorphic enzyme activityconverting alpha-ketoglutarate to 2-hydroxyglutarate. Cancer Cell17, 225234

28 Pardanani, A. et al. (2010) IDH1 and IDH2 mutation analysis inchronic- and blast-phase myeloproliferative neoplasms. Leukemia 24,11461151

29 Abbas, S. et al. (2010) Acquired mutations in the genes encoding IDH1and IDH2 both are recurrent aberrations in acute myeloid leukemia:prevalence and prognostic value. Blood 116, 21222126

30 Paschka, P. et al. (2010) IDH1 and IDH2 mutations are frequentgenetic alterations in acute myeloid leukemia and confer adverseprognosis in cytogenetically normal acute myeloid leukemia withNPM1 mutation without FLT3 internal tandem duplication.J. Clin. Oncol. 28, 36363643

31 Yan, H. et al. (2009) IDH1 and IDH2 mutations in gliomas. N. Engl. J.Med. 360, 765773

32 Koivunen, P. et al. (2012) Transformation by the (R)-enantiomer of2-hydroxyglutarate linked to EGLN activation. Nature 483, 484488

33 Bertout, J.A. et al. (2009) Heterozygosity for hypoxia inducible factor1alpha decreases the incidence of thymic lymphomas in a p53 mutantmouse model. Cancer Res. 69, 32133220

34 Lidgren, A. et al. (2005) The expression of hypoxia-inducible factor1alpha is a favorable independent prognostic factor in renal cellcarcinoma. Clin. Cancer Res. 11, 11291135

35 Klatte, T. et al. (2007) Hypoxia-inducible factor 1 alpha in clear cellrenal cell carcinoma. Clin. Cancer Res. 13, 73887393

36 Mazumdar, J. et al. (2010) HIF-2alpha deletion promotes Kras-drivenlung tumor development. Proc. Natl. Acad. Sci. U.S.A. 107,

141821418737 Velasco-Hernandez, T. et al. (2014) HIF-1alpha can act as atumor suppressor gene in murine acute myeloid leukemia. Blood 124,35973607

38 Zhang, H. et al. (2007) HIF-1 inhibits mitochondrial biogenesis andcellular respiration in VHL-deficient renal cell carcinoma byrepression of C-MYC activity. Cancer Cell 11, 407420

39 Mazure, N.M. et al. (1996) Oncogenic transformation and hypoxiasynergistically act to modulate vascular endothelial growth factorexpression. Cancer Res. 56, 34363440

40 Arany, Z. et al. (1996) An essential role for p300/CBP in thecellular response to hypoxia. Proc. Natl. Acad. Sci. U.S.A. 93,1296912973

41 Forsythe, J.A. et al. (1996) Activation of vascular endothelial growthfactor gene transcription by hypoxia-inducible factor 1. Mol. Cell. Biol.16, 46044613

42 Kong, D. et al. (2005) Echinomycin, a small-molecule inhibitor ofhypoxia-inducible factor-1 DNA-binding activity. Cancer Res. 65,90479055

43 Comerford, K.M. et al. (2002) Hypoxia-inducible factor-1-dependentregulation of the multidrug resistance (MDR1) gene. Cancer Res. 62,33873394

44 Ricker, J.L. et al. (2004) 2-methoxyestradiol inhibits hypoxia-inducible factor 1alpha, tumor growth, and angiogenesis andaugments paclitaxel efficacy in head and neck squamous cellcarcinoma. Clin. Cancer Res. 10, 86658673

45 Zhang, H. et al. (2014) Kruppel-like factor 8 contributes to hypoxia-induced MDR in gastric cancer cells. Cancer Sci. 105, 11091115

46 Ria, R. et al. (2014) HIF-1alpha of bone marrow endothelial cellsimplies relapse and drug resistance in patients with multiplemyeloma and may act as a therapeutic target. Clin. Cancer Res.20, 847858

47 Li, D.W. et al. (2013) Hypoxia induced multidrug resistance oflaryngeal cancer cells via hypoxia-inducible factor-1alpha. AsianPac. J. Cancer Prev. 14, 48534858

48 Wang, H. et al. (2014) Wogonin reverses hypoxia resistance of humancolon cancer HCT116 cells via downregulation of HIF-1alpha andglycolysis, by inhibiting PI3K/Akt signaling pathway. Mol. Carcinog.53 (Suppl. 1), E107E118

49 Cui, X.Y. et al. (2013) Hypoxia influences stem cell-like properties inmultidrug resistant K562 leukemic cells. Blood Cells. Mol. Dis. 51,177184

50 Ji, Z. et al. (2010) Expression of MDR1, HIF-1alpha and MRP1 insacral chordoma and chordoma cell line CM319. J. Exp. Clin. CancerRes. 29, 158

51 Ding, Z. et al. (2010) Expression and significance of hypoxia-induciblefactor-1 alpha and MDR1/P-glycoprotein in human colon carcinomatissue and cells. J. Cancer Res. Clin. Oncol. 136, 16971707

52 An, W.G. et al. (1998) Stabilization of wild-type p53 by hypoxia-inducible factor 1alpha. Nature 392, 405408

53 Shoshani, T. et al. (2002) Identification of a novel hypoxia-induciblefactor 1-responsive gene, RTP801, involved in apoptosis. Mol. Cell.Biol. 22, 22832293

54 Bruick, R.K. (2000) Expression of the gene encoding the proapoptoticNip3 protein is induced by hypoxia. Proc. Natl. Acad. Sci. U.S.A. 97,90829087

55 Sowter, H.M. et al. (2001) HIF-1-dependent regulation of hypoxicinduction of the cell death factors BNIP3 and NIX in humantumors. Cancer Res. 61, 66696673

56 Koshiji, M. et al. (2004) HIF-1alpha induces cell cycle arrest byfunctionally counteracting Myc. EMBO J. 23, 19491956

57 Jin, L. et al. (2006) Targeting of CD44 eradicates human acutemyeloid leukemic stem cells. Nat. Med. 12, 11671174

58 Jin, L. et al. (2009) Monoclonal antibody-mediated targeting ofCD123, IL-3 receptor alpha chain, eliminates human acute myeloidleukemic stem cells. Cell Stem Cell 5, 3142

59 Kikushige, Y. et al. (2010) TIM-3 is a promising target to selectivelykill acute myeloid leukemia stem cells. Cell Stem Cell 7, 708717

60 Guzman, M.L. et al. (2007) An orally bioavailable parthenolide analogselectively eradicates acute myelogenous leukemia stem andprogenitor cells. Blood 110, 44274435

61 Kelly, P.N. et al. (2007) Tumor growth need not be driven by rare

cancer stem cells. Science 317, 337

381

Review Trends in Pharmacological Sciences June 2015, Vol. 36, No. 662 Wang, L. et al. (2013) FISH+CD34+CD38 cells detected in newlydiagnosed acute myeloid leukemia patients can predict the clinicaloutcome. J. Hematol. Oncol. 6, 85

63 Witte, K.E. et al. (2011) High proportion of leukemic stem cells atdiagnosis is correlated with unfavorable prognosis in childhood acutemyeloid leukemia. Pediatr. Hematol. Oncol. 28, 9199

64 van Rhenen, A. et al. (2005) High stem cell frequency in acute myeloidleukemia at diagnosis predicts high minimal residual disease andpoor survival. Clin. Cancer Res. 11, 65206527

65 Gentles, A.J. et al. (2010) Association of a leukemic stem cell geneexpression signature with clinical outcomes in acute myeloidleukemia. JAMA 304, 27062715

66 (2014) NIH Clinical Center open to coll-aboration. Cancer Discovery 4,752

67 Wang, Y. et al. (2014) Echinomycin protects mice against relapsedacute myeloid leukemia without adverse effect on hematopoietic stemcells. Blood 124, 11271135

68 Bernot, K.M. et al. (2013) Toward personalized therapy in AML: invivo benefit of targeting aberrant epigenetics in MLL-PTD-associatedAML. Leukemia 27, 23792382

69 Sloma, I. et al. (2010) Insights into the stem cells of chronic myeloidleukemia. Leukemia 24, 18231833

70 Mayerhofer, M. et al. (2002) BCR/ABL induces expression of vascularendothelial growth factor and its transcriptional activator, hypoxiainducible factor-1alpha, through a pathway involving phosphoinositide3-kinase and the mammalian target of rapamycin. Blood 100,37673775

71 Zhao, F. et al. (2010) Imatinib resistance associated with BCR-ABLupregulation is dependent on HIF-1alpha-induced metabolicreprograming. Oncogene 29, 29622972

72 Zhang, H. et al. (2012) HIF1alpha is required for survivalmaintenance of chronic myeloid leukemia stem cells. Blood 119,25952607

73 Ng, K.P. et al. (2014) Physiologic hypoxia promotes maintenance ofCML stem cells despite effective BCR-ABL1 inhibition. Blood 123,33163326

74 Conley, S.J. et al. (2012) Antiangiogenic agents increase breast cancerstem cells via the generation of tumor hypoxia. Proc. Natl. Acad. Sci.U.S.A. 109, 27842789

75 Schwab, L.P. et al. (2012) Hypoxia-inducible factor 1alpha promotesprimary tumor growth and tumorinitiating cell activity in breastcancer. Breast Cancer Res. 14, R6

76 Cordenonsi, M. et al. (2011) The Hippo transducer TAZ conferscancer stem cell-related traits on breast cancer cells. Cell 147,759772

77 Xiang, L. et al. (2014) Hypoxia-inducible factor 1 mediates TAZexpression and nuclear localization to induce the breast cancerstem cell phenotype. Oncotarget 5, 1250912527

78 Chaturvedi, P. et al. (2014) Hypoxia-inducible factor-dependentsignaling between triple-negative breast cancer cells andmesenchymal stem cells promotes macrophage recruitment. Proc.Natl. Acad. Sci. U.S.A. 111, E2120E2129

79 Lukashev, D. et al. (2001) Differential regulation of two alternativelyspliced isoforms of hypoxia-inducible factor-1 alpha in activatedT lymphocytes. J. Biol. Chem. 276, 4875448763

80 Lukashev, D. et al. (2006) Cutting edge: hypoxia-inducible factor1alpha and its activation-inducible short isoform I.1 negativelyregulate functions of CD4+ and CD8+ T lymphocytes. J. Immunol.177, 49624965

81 Dang, E.V. et al. (2011) Control of T(H)17/T(reg) balance by hypoxia-inducible factor 1. Cell 146, 772784

82 Shi, L.Z. et al. (2011) HIF1alpha-dependent glycolytic pathwayorchestrates a metabolic checkpoint for the differentiation of TH17and Treg cells. J. Exp. Med. 208, 13671376

83 Kryczek, I. et al. (2011) Human TH17 cells are longlived effectormemory cells. Sci. Transl. Med. 3, 104ra100

84 Ben-Shoshan, J. et al. (2008) Hypoxia controls CD4+CD25+regulatory T-cell homeostasis via hypoxia-inducible factor-1alpha.Eur. J. Immunol. 38, 24122418

85 Facciabene, A. et al. (2011) Tumour hypoxia promotes toleranceand angiogenesis via CCL28 and T(reg) cells. Nature 475,

226230

38286 Finlay, D.K. et al. (2012) PDK1 regulation of mTOR and hypoxia-inducible factor 1 integrate metabolism and migration of CD8+ T cells.J. Exp. Med. 209, 24412453

87 Doedens, A.L. et al. (2013) Hypoxia-inducible factors enhance theeffector responses of CD8(+) T cells to persistent antigen. Nat.Immunol. 14, 11731182

88 Thiel, M. et al. (2007) Targeted deletion of HIF1alpha gene in T cellsprevents their inhibition in hypoxic inflamed tissues and improvesseptic mice survival. PLoS ONE 2, e853

89 Jantsch, J. et al. (2008) Hypoxia and hypoxia-inducible factor-1 alphamodulate lipopolysaccharide-induced dendritic cell activation andfunction. J. Immunol. 180, 46974705

90 Noman, M.Z. et al. (2014) PD-L1 is a novel direct target of HIF-1alpha,and its blockade under hypoxia enhanced MDSC-mediated T cellactivation. J. Exp. Med. 211, 781790

91 Barsoum, I.B. et al. (2014) A mechanism of hypoxia-mediated escapefrom adaptive immunity in cancer cells. Cancer Res. 74, 665674

92 Curiel, T.J. et al. (2003) Blockade of B7-H1 improves myeloiddendritic cell-mediated antitumor immunity. Nat. Med. 9, 562567

93 Gabrilovich, D.I. and Nagaraj, S. (2009) Myeloid-derived suppressorcells as regulators of the immune system. Nat. Rev. Immunol. 9, 162174

94 Talmadge, J.E. and Gabrilovich, D.I. (2013) History of myeloid-derived suppressor cells. Nat. Rev. Cancer 13, 739752

95 Corzo, C.A. et al. (2010) HIF-1alpha regulates function anddifferentiation of myeloid-derived suppressor cells in the tumormicroenvironment. J. Exp. Med. 207, 24392453

96 Doedens, A.L. et al. (2010) Macrophage expression of hypoxia-inducible factor-1 alpha suppresses T-cell function and promotestumor progression. Cancer Res. 70, 74657475

97 Chen, W. et al. (2010) mTOR signaling is activated by FLT3 kinaseand promotes survival of FLT3mutated acute myeloid leukemiacells. Mol. Cancer 9, 292

98 Gilliland, D.G. and Griffin, J.D. (2002) The roles of FLT3 inhematopoiesis and leukemia. Blood 100, 15321542

99 Sathaliyawala, T. et al. (2010) Mammalian target of rapamycincontrols dendritic cell development downstream of Flt3 ligandsignaling. Immunity 33, 597606

100 Brugarolas, J. et al. (2004) Regulation of mTOR function in responseto hypoxia by REDD1 and the TSC1/TSC2 tumor suppressor complex.Genes Dev. 18, 28932904

101 Brugarolas, J.B. et al. (2003) TSC2 regulates VEGF through mTOR-dependent and -independent pathways. Cancer Cell 4, 147158

102 Hudson, C.C. et al. (2002) Regulation of hypoxia-inducible factor1alpha expression and function by the mammalian target ofrapamycin. Mol. Cell. Biol. 22, 70047014

103 Van de Velde, S. et al. (2011) mTOR links incretin signaling to HIFinduction in pancreatic beta cells. Proc. Natl. Acad. Sci. U.S.A. 108,1687616882

104 Huang, J. et al. (2009) Pivotal role for glycogen synthase kinase-3 inhematopoietic stem cell homeostasis in mice. J. Clin. Invest. 119,35193529

105 Parkin, B. et al. (2010) Acquired genomic copy number aberrations andsurvival in adult acute myelogenous leukemia. Blood 116, 49584967

106 Joshi, S. et al. (2014) MDM2 regulates hypoxic hypoxia-induciblefactor 1alpha stability in an E3 ligase, proteasome, and PTEN-phosphatidylinositol 3-kinase-AKT-dependent manner. J. Biol.Chem. 289, 2278522797

107 Ravi, R. et al. (2000) Regulation of tumor angiogenesis by p53-induceddegradation of hypoxia-inducible factor 1alpha. Genes Dev. 14, 3444

108 Falini, B. et al. (2005) Cytoplasmic nucleophosmin in acute myelogenousleukemia with a normal karyotype. N. Engl. J. Med. 352, 254266

109 Grisendi, S. and Pandolfi, P.P. (2005) NPM mutations in acutemyelogenous leukemia. N. Engl. J. Med. 352, 291292

110 Chen, D. et al. (2010) Transcription-independent ARF regulation inoncogenic stress-mediated p53 responses. Nature 464, 624627

111 Fatyol, K. and Szalay, A.A. (2001) The p14ARF tumor suppressorprotein facilitates nucleolar sequestration of hypoxia-inducible factor-1alpha (HIF-1alpha) and inhibits HIF-1-mediated transcription.J. Biol. Chem. 276, 2842128429

112 Wang, H. et al. (2014) Negative regulation of Hif1a expression andTH17 differentiation by the hypoxia-regulated microRNA miR-210.

Nat. Immunol. 15, 393401

113 Blouin, C.C. et al. (2004) Hypoxic gene activation by lipopolysaccharidein macrophages: implication of hypoxia-inducible factor 1alpha. Blood103, 11241130

114 Laughner, E. et al. (2001) HER2 (neu) signaling increases the rate ofhypoxia-inducible factor 1alpha (HIF-1alpha) synthesis: novelmechanism for HIF-1-mediated vascular endothelial growth factorexpression. Mol. Cell. Biol. 21, 39954004

115 Cui, T.X. et al. (2013) Myeloid-derived suppressor cells enhancestemness of cancer cells by inducing microRNA101 and suppressingthe corepressor CtBP2. Immunity 39, 611621

116 Kryczek, I. et al. (2014) IL-22(+)CD4(+) T cells promote colorectalcancer stemness via STAT3 transcription factor activation andinduction of the methyltransferase DOT1L. Immunity 40, 772784

117 Lin, E.Y. et al. (2006) Macrophages regulate the angiogenic switch in amouse model of breast cancer. Cancer Res. 66, 1123811246

118 Ben-Neriah, Y. and Karin, M. (2011) Inflammation meets cancer, withNF-kappaB as the matchmaker. Nat. Immunol. 12, 715723

119 Lin, E.Y. and Pollard, J.W. (2007) Tumor-associated macrophagespress the angiogenic switch in breast cancer. Cancer Res. 67,50645066

Review Trends in Pharmacological Sciences June 2015, Vol. 36, No. 6383

Hypoxia-inducible factors in cancer stem cells and inflammationChallenges in targeting the tumor microenvironmentHIF and cancer: an overviewPKM2 and metabolic switches in cancer cellsVEGF and cancer neoangiogenesisMDR1Tumor growth inhibition by HIF-1

HIF and CSCsGlioma stem cells and HIFAcute lymphocytic leukemiaAMLChronic myeloid leukemiaBreast CSCs

HIF in adaptive and innate inflammationTreg/TH17 cell differentiationCTL differentiation and functionDendritic cell activationTumor-associated myeloid suppressor cells

Epigenetic and genetic mechanisms for HIF activities under normoxia: cross-fertilization between cancer biology and immuno...Genetic mechanisms taught by cancer: the AML example

Epigenetic mechanism of HIF1a activation: unanswered questions from CSCs and immune effector cellsConcluding remarksAcknowledgmentsReferences