A Macrophage-Dominant PI3K Isoform Controls …...Shweta Joshi 1, Alok R. Singh , Muamera Zulcic , and Donald L. Durden1,2,3 Abstract Tumor growth, progression, and response to the

Signal Transduction

A Macrophage-Dominant PI3K Isoform Controls Hypoxia-Induced HIF1a and HIF2a Stability and Tumor Growth,Angiogenesis, and Metastasis

Shweta Joshi1, Alok R. Singh1, Muamera Zulcic1, and Donald L. Durden1,2,3

AbstractTumor growth, progression, and response to the hypoxic tumor microenvironment involve the action of

hypoxia-inducible transcription factors, HIF1 andHIF2. HIF is a heterodimeric transcription factor containing aninducible HIFa subunit and a constitutively expressed HIFb subunit. The signaling pathways operational inmacrophages regulating hypoxia-induced HIFa stabilization remain the subject of intense investigation. Here, itwas discovered that the PTEN/PI3K/AKT signaling axis controls hypoxia-induced HIF1a (HIF1A) and HIF2a(EPAS1) stability inmacrophages. Using geneticmousemodels and pan-PI3K as well as isoform-specific inhibitors,inhibition of the PI3K/AKT pathway blocked the accumulation of HIFa protein and its primary transcriptionaltarget VEGF in response to hypoxia. Moreover, blocking the PI3K/AKT signaling axis promoted the hypoxicdegradation of HIFa via the 26S proteasome. Mechanistically, a macrophage-dominant PI3K isoform (p110g)directed tumor growth, angiogenesis, metastasis, and the HIFa/VEGF axis. Moreover, a pan-PI3K inhibitor(SF1126) blocked tumor-induced angiogenesis and inhibited VEGF and other proangiogenic factors secreted bymacrophages. These data define a novel molecular mechanism by which PTEN/PI3K/AKT regulates theproteasome-dependent stability of HIFa under hypoxic conditions, a signaling pathway in macrophages thatcontrols tumor-induced angiogenesis and metastasis.

Implications: This study indicates that PI3K inhibitors are excellent candidates for the treatment of cancers wheremacrophages promote tumor progression. Mol Cancer Res; 12(10); 1520–31. �2014 AACR.

IntroductionHypoxia is considered a hallmark of tumor progression in

that it induces neoangiogenesis, the recruitment of bonemarrow–derived cells, and degradation of extracellularmatrix (ECM) resulting in activation of proliferation, inva-sion, and metastasis of cancer cells (1). Hypoxic regions oftumors are characterized by an increased accumulation ofmacrophages which contributes to tumor angiogenesis andtumor progression (2). Hypoxic macrophages are knownto respond to hypoxia by upregulating hypoxia-inducibletranscription factors (HIF) consisting of distinct, hypoxia-responsive (HIF1a and HIF2a) subunits and identical,constitutively expressed (HIF1b and HIF2b) subunits

(3, 4). In the presence of oxygen, thea subunits are hydroxy-lated by oxygen-sensitive enzymes, prolyl hydroxylases,which targets them for degradation by the von Hippel-Lindau (VHL)-dependent ubiquitin–proteasomal pathway(5, 6). Under hypoxic conditions, HIFa subunits are sta-bilized, translocate to the nucleus, and, together with theirpartner factor, basic-helix-loop-helix/PAS protein Arnt,bind to the promoters of genes that mediate glycolysis andneovascularization/angiogenesis (7, 8).There are conflicting views of the relative contribution of

each HIF to the regulation of hypoxic gene expression inmacrophages. Some studies suggest that the main form ofHIF upregulated by tumor-associated macrophages (TAM)is HIF2 (9, 10), and overexpression of HIF2a in normoxichuman macrophages upregulates various proangiogenicgenes (11). However, human macrophages also markedlyupregulate HIF1a when exposed to hypoxia in vitro and intumors (12), and HIF1a-deficient murine macrophagesexpress lower levels of such HIF-regulated genes such asVEGF and the glucose receptor GLUT1 in hypoxia thantheir wild-type (WT) counterparts. Interestingly, the exactcontribution of HIF1a and HIF2a to the regulation ofhypoxic gene expression appears to vary between differentcell types (7).The molecular mechanisms controlling HIFa accumula-

tion in macrophages are not completely elucidated. In the

1UCSD Department of Pediatrics, Moores Cancer Center, University ofCalifornia, La Jolla, California. 2Division of Pediatric Hematology-Oncolo-gy, UCSD Rady Children's Hospital, San Diego, California. 3SignalRxPharmaceuticals, San Diego, California.

Note: Supplementary data for this article are available at Molecular CancerResearch Online (http://mcr.aacrjournals.org/).

Corresponding Author: Donald L. Durden, UCSD Department of Pediat-rics,MooresUCSDCancerCenter, 3855Health SciencesDrive, SanDiego,CA, 92093. Phone: 858-534-3355; Fax: 858-822-0022; E-mail:[email protected]

doi: 10.1158/1541-7786.MCR-13-0682

�2014 American Association for Cancer Research.

MolecularCancer

Research

Mol Cancer Res; 12(10) October 20141520

on June 12, 2020. © 2014 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst August 7, 2014; DOI: 10.1158/1541-7786.MCR-13-0682

http://mcr.aacrjournals.org/

tumor cells, a large number of signaling pathways, includingRac-GTP, Src, Fak, PI3K, etc., have been linked to thecontrol of HIF1a and VEGF (13, 14). Several groups haveimplicated the role of PTEN/PI3K signaling in the regula-tion of HIF1a (15–17) in tumor cells. However, the exactcontribution of the PI3K pathway to HIFa regulationremains a subject of considerable interest and controversyin both the tumor and stromal compartments (18, 19).PI3K are a family of enzymes that phosphorylate phos-

phatidylinositol (PI) lipids at the 30 position (20). ThePTEN deleted on chromosome 10 (PTEN) is a dual spec-ificity phosphatase that antagonizes the enzymatic activity ofPI3K by dephosphorylating phosphatidylinositol (3,4,5)-trisphosphate to generate phosphatidylinositol (4,5)-bipho-sphate. PI3K/PTEN has been shown to play a prominentrole in regulating a number of cellular responses, includinggrowth, survival, and migration (21, 22). PI3K is known tobe associated with different phases of macrophage function(23). The PI3K/Akt signaling axis is crucial for migrationand diapedesis of immune effector cells, such as neutrophilsand monocytes, to the site of inflammation and infection(24–26). The role of myeloid-specific PTEN in controllinglung infection and inflammation during murine pneumo-coccal pneumonia and neutropenia-associated pneumonia isalso well-studied (27, 28).However, the contribution of PTEN/PI3K/AKTpathway

in enhancing HIFa-mediated gene expression in macro-phages has not been previously reported. In the presentstudy, using myeloid-specific deletion of PTEN and variouspan and isoform-specific PI3K inhibitors, we show thatPTEN/PI3K signaling controls hypoxia-induced HIF1aand HIF2a stabilization in macrophages. Moreover, usingp110g�/� mice and pan-PI3K inhibitor SF1126, we dem-onstrate that PI3K signaling promotes tumor growth andmetastasis in tumor models by increasing the expression andpotential secretion of proangiogenic factors by TAMs.

Materials and MethodsAnimal studiesAll procedures involving animals were approved by the

University of California SanDiego Animal CareCommittee,which serves to ensure that all federal guidelines concerninganimal experimentation are met. PI3K g (p110g�/�) micewere obtained from Dr. Joseph Penninger, Institute ofMolecular Biotechnology (Vienna, Austria). Floxed PTENmice and lysozyme M (LysM) Cre recombinase transgenicmice were purchased from Jackson laboratories. LysM Cremice were mated with PTENfl/fl mice to delete PTEN frommyeloid cell compartment. Depending on the presence ofLysM Cre recombinase, double-floxed PTENfl/fl mice arereferred to as PTENMC-KO or PTENMC-Wt mice, respec-tively. Macrophages from PTENfl/fl mice were used as WTcontrols.

Antibodies and reagentsMacrophage colony-stimulating factor (MCSF) was from

Gibco Life technologies. PI-103, BKM120, and PF4691502were purchased from Selleck chemicals. SF2523 and SF1126

were obtained from SignalRx Pharmaceuticals (29, 30).Primary or fluorescent antibodies against CD31 (cloneMEC13.3) and CD11b (clone M1/70) were from BDBiosciences; F4/80 (clone BM8) was from eBiosciences.Alexa Flour 488 was from Invitrogen life Technologies.Bouin's solution, 4,60-diamidino-2-phenylindole (DAPI),hyaluronidase type V, and DNase I were from Sigma.

Isolation of bonemarrow–derivedmacrophages, hypoxiaand in vitro inhibitor experimentsBone marrow–derived macrophages (BMDM) were iso-

lated as described previously and cultured under MCSF (75ng/mL)þ 20% FBS conditions (31). On day 7, most of theadherent cells were macrophages as confirmed by FACSanalysis (>90% Mac1- and F4/80-positive cells). To studythe effect of different inhibitors in vitro, macrophages (7 daysin culture) were treated with or without pan or isoform-specific PI3K inhibitors followed by hypoxia (1% O2) for 4hours. For hypoxia experiments,macrophages were placed ina modulator incubator chamber (Billups-Rothenberg, Inc.)that was flushed with a gas mixture consisting of 1%O2, 5%CO2, with balance N2, sealed, and incubated at 37�C.Whole-cell lysates were prepared in RIPA buffer (CellSignaling Technologies) containing complete proteaseinhibitor cocktail. For HIFa immunoblots, nuclear extractswere prepared using Dignam protocol (32).

Plasmid and siRNA transfectionsBMDMs were transfected using an AMAXA Mouse

Macrophage Nucleofection Kit (Lonza) with 5 mg of HA-PTEN or myrAKT plasmid or 100 nmol/L of siRNA forPI3Ka (Mm_pik3ca_1), PI3Kb (Mm_pik3cb_2), PI3Kg(Mm_pik3cg_1), PI3Kd (Mm_pik3cd_1), Akt1/2 siRNAor nonsilencing siRNA (Ctrl_AllStars-1) purchased fromQiagen or Akt1/2 siRNA and control siRNA purchasedfrom Santa Cruz Biotechnology. After transfection, cellswere cultured for 36 hours. Each siRNA was tested indi-vidually for efficient knockdown of protein expression.

Quantification of gene expressionTotal RNAwas isolated fromBMDMs using RNAeasy kit

(Qiagen). cDNAwas prepared from 1 mg RNA sample usingiScript cDNA synthesis kit (Bio-Rad). cDNA (2 mL) wasamplified by RT-PCR reactions with 1� SYBR green super-mix (Bio-) in 96-well plates on an CFX96 Real time system(Bio-Rad), using different primers formouse gene. Sequenceof primers used were reported before (31). Relative expres-sion levels were normalized toGAPDHexpression accordingto the formula: 2Ct_gene of interest � Ct_GAPDH. Values aremultiplied by 100 for presentation purposes.

ImmunoblottingWhole-cell extract (WCE) or nuclear extracts of BMDMs

were prepared and protein was quantitated with Bio-Radprotein assay kit (Bio- Rad Laboratories) using BSA as astandard. Equal protein fromprotein lysates were resolved bySDS-PAGE, followed by immunoblotting and probingwith primary antibodies against p110a (Cell Signaling

PTEN/PI3K Regulates HIFa/VEGF Axis and Metastasis in Macrophages

www.aacrjournals.org Mol Cancer Res; 12(10) October 2014 1521




Technologies), p110b (Santa Cruz), p110g (Cell SignalingTechnologies) and p110d (Santa Cruz), pAKT (Cell Sig-naling), AKT (Cell Signaling), HIF1a and HIF2a (NovusBiological), and b-actin (Santa Cruz).

In vivo tumor experimentsLewis lung carcinoma (LLC) and B16 F10 melanoma

obtained from the ATCC were cultured as described pre-viously (31). LLC cells (1 � 105) were injected subcutane-ously into syngeneic 4- to 6-week-old mice. Tumor dimen-sions were recorded regularly, and tumors were harvested 25days postinjection. Tumor volume was measured using thefollowing formula: Volume¼ 0.5� length� (width)2. Forexperimental metastasis, B16 F10 melanoma cells (5� 105)were injected intravenously, and lungs were harvested after15 days as described before (31). For drug treatment studies,WT mice implanted subcutaneously with 1 � 105 LLCcells were treated with vehicle or 50 mg/kg SF1126 subcu-taneously 3 times a week for 28 days, until tumors wereharvested. For B16 F10 melanoma, 2 doses of 50 mg/kgof SF1126 were given before injecting 5 � 105 B16F10melanoma cells, followed by daily treatment with the druguntil lungs were harvested on day 15. For survival studies,3 doses of 50 mg/kg of SF1126 were given before injecting

3 � 105 B16F10 melanoma cells, followed by daily treat-ment until mice died. Tumor volumes and weights weremeasured and the tumors were used for immunofluores-cence, IHC, and sorting of macrophages using FACS. ForFACS analysis, LLC tumors were excised, minced, anddigested to single-cell population as described (31). CD11b-and F4/80-positive cells sorted by FACS were used for RNAisolation and real-time PCR studies.

ResultsPTEN regulates expression of HIFa protein and HIFamRNA targets in macrophagesThe tumor suppressor PTEN is an antagonist of PI3K

signaling and functions by removing the phosphate at theD3 position of phosphatidylinositol trisphosphate and phos-phatidylinositol bisphosphate. To validate the role of PTENin hypoxia-induced HIFa stabilization, we generated con-ditional PTEN knockout mice in which PTEN expressionwas controlled by LysM, a myeloid cell–specific promoter(20). Depending on the presence of LysM Cre recombinase,double-floxed PTENfl/flmice are referred to as PTENMC-KO

or PTENMC-Wt mice, respectively. Deletion of PTENwas confirmed in primary macrophages (Fig. 1A), andresulting downstream effects of constitutively active PI3K

PTENMC-WT

A B

C D E

85 90 86 91 Mouse ID 85 90 86 91 Mouse ID

HIF1a

HIF2a

HIF1a

HIF2a

PTEN

b-Actin

HIF1a

HIF2a

HA-PTEN

HA-PTEN– +Normoxia Hypoxia

– +

b-Actin

PTEN

pAKT

AKT

b-Actin

50

40

30

20

10

0b-Actin

Mouse ID

PTEN+/+

211 212 204 205

PTEN+/–

PTENMC-ko PTENMC-WT

PTENMC-ko Mice

PTENMC-ko

PTENMC-WTRel

ativ

e V

EG

F e

xpre

ssio

n

PTENMC-ko

Figure 1. Deletion of PTEN in myeloid cells increase hypoxic HIF1a and HIF2a accumulation. A, Western blot analysis showing deletion of PTEN in myeloidcells. Cell lysates from BMDMs isolated from PTENMC-KO or PTENMC-Wt mice were subjected toWestern blot analysis for PTEN, pAKT, and AKT. b-Actin wasused as loading control. B and C, BMDMs isolated from PTENMC-WT and PTENMC-KO (B) and PTENþ/þ and PTENþ/� (C) mice were incubated under hypoxic(1%O2) conditions for 4 hours followedby preparation of nuclear extracts (for HIF1a andHIF2a) andWestern blot analysis. Each lane in A, B, andC representslysate prepared from individual mice. Mouse ID numbers are provided for each genotype. D, mRNA was isolated from PTENMC-WT and PTENMC-KO

BMDMs incubated under hypoxia (1% O2) for 24 hours. VEGF mRNA expression was measured by real-time PCR. Data represent mean � SEM (n ¼ 3 or 4;P < 0.001; pairwise 2-sided Student t test). E, BMDMs isolated from PTENMC-KO mice were transfected with 5 mg of HA-PTEN, using AMAXA mousemacrophageNucleofaction kit. After 36 hours of transfection, cells were exposed to normoxia (21%O2) or hypoxia (1%O2) for 4 hours followed by preparationof either nuclear extracts (for HIFa blots) orWCE (PTENblot) andWestern blot analysis. Experimentswere repeated 4 to 5 timeswith 2 to 3mice in each group.

Joshi et al.

Mol Cancer Res; 12(10) October 2014 Molecular Cancer Research1522




were reflected by greatly elevated baseline levels of phospho-Akt in PTENMC-KO macrophages (Fig. 1A). Myeloid-spe-cific deletion of PTEN induced high levels of HIF1a andHIF2a stabilization in PTENMC-KOmice as compared withPTENMC-Wt (Fig. 1B). These results were also supported bya greater degree of HIFa stabilization in BMDMs isolatedfrom PTENþ/�mice (Fig. 1C). Our next point of focus wasto investigate the role of PTEN in the transcriptionalactivation of the HIFa target gene VEGF. Loss of PTENin myeloid cell compartment increased VEGF mRNA levelsin PTENMC-KO mice (Fig. 1D). To further validate the roleof PTEN in stabilization of HIFa, PTEN was exogenouslyexpressed in PTENMC-ko mice. Results in Fig. 1E showedthat expression of PTENblocked the accumulation ofHIFain PTENMC-ko mice. Taken together, these results suggest arole for PTEN in the regulation of the HIFa/VEGF axisunder hypoxia in macrophages (Fig. 1).

Inhibition of PI3K/AKT pathway blocks the hypoxicinduction of HIF1a and HIF2a and its transcriptionaltarget VEGF in macrophagesThe observation that PTEN exerts control over HIF1a

and HIF2a protein expression and transcriptional activity(Fig. 1) led us to investigate the effects of different clinicallyrelevant PI3K inhibitors (SF1126, PF4691502, PI-103, andBKM120) on the hypoxic induction of HIFa in macro-phages. SF1126, a pan-PI3K inhibitor, is a vascular RGDS/integrin targeted prodrug derivative of LY294002 and isentering phase II clinical trials (29). The other inhibitorsused in early-phase clinical trials include: PI-103, a dualPI3K/mTOR inhibitor; PF4691502, a pan-PI3K inhibitor;and BKM120, a pan-PI3K inhibitor (33–35). The resultsshown in Fig. 2A clearly demonstrate that SF1126, PI-103,PF4691502, and BKM120 potently inhibit the hypoxicinduction of HIF1a and HIF2a in BMDMs. Furthermore,

Co

ntr

ol

500

nm

ol/L

PF

4691

502

500

nm

ol/L

PI-

103

500

nm

ol/L

BK

M12

0

25 m

mo

l/L S

F11

26

Co

ntr

ol

10 m

mo

l/L S

F25

23

20 m

mo

l/L S

F25

23

50 m

mo

l/L S

F25

23

10 mm

ol/L S

F2523

20 mm

ol/L S

F2523

50 mm

ol/L S

F2523

Rel

ativ

e V

EG

F e

xpre

ssio

n

Rel

ativ

e V

EG

F e

xpre

ssio

n

HIF1aHIF2a

b-Actin

pAKTAKT

HIF1a

HIF2a

b-Actin

pAKT

AKT

50A B C

D E F

40

30

20

10

0

50

40

30

20

10

0

Control

Control

500 n

mol/L

PF46

9150

2

500 n

mol/L

PI-1

03

500 n

mol/L

BKM12

0

25 mm

ol/L S

F1126

Ctrl AKT siRNA myr-AKT– +

HIF1a

HIF2a

AKT

b-Actin

HIF1a

HIF2a

AKT

b-Actin

Figure 2. PI3K/AKT signaling axis control HIFa/VEGF axis. A and B, BMDMs from WT animals were treated with different concentrations of PI3K inhibitors(500 nmol/L PF4691502, PI-103, BKM120 and 25 mmol/L SF1126) followed by hypoxia for 4 hours for Western blots or 24 hours for real-time PCR studies.These macrophages were either used for lysate preparation (nuclear extracts for HIFa or WCE for pAKT and AKT) and Western blot analysis (A) orRNA and real-time PCR for VEGF (B). C and D,WT BMDMs were treated with different concentrations of SF2523 (10, 20, and 50 mmol/L) followed by hypoxiafor 4 hours for Western blot analyses (C) or 24 hours for RNA isolation and real-time PCR for VEGF (D). E and F, BMDMs from WT animals weretransfected with either 5 mg of myrAKT plasmid or 100 nmol/L control siRNA or 100 nmol/L siRNA against AKT1/2 using Amaxa Mouse MacrophageNucleofaction kit (Lonza). Thirty-six hours after transfection, hypoxia was given to these transfected BMDMs for 4 hours followed by cell lysate preparationandWestern blot analysis. Graphs in B and D represent mean�SEM (n¼ 3–4). Statistical significance is assessed by 2-sample t test. �,P < 0.05; ��,P < 0.01;��, P < 0.001. Experiments were repeated 3 to 4 times with similar results.






these pan-PI3K inhibitors also blocked the transcriptionalactivation of HIFa target gene, VEGF (Fig. 2B). In additionto pan-PI3K inhibitors in clinical trial, we used SF2523, anovel pan-PI3K inhibitor developed in collaboration withSignalRx Pharmaceuticals (30), to see the effect of this PI3Kinhibitor on HIFa stabilization. SF2523 blocked accumu-lation of HIF1a and HIF2a in dose-dependent manner(Fig. 2C), which correlates with the decreasedVEGFmRNAin BMDMs treated with SF2523 (Fig. 2D). To determinethe role of AKT in hypoxic HIFa accumulation, experi-ments were performed where AKT is either knocked downor an activated membrane-targeted AKT (myrAKT) isexogenously expressed in BMDMs. Figure 2E and F showedthat knockdown of AKT using siRNA blocked accumula-tion of HIFa, whereas exogenous expression of myrAKTincreased HIFa levels.

Specific p110d and p110 g isoforms of PI3K controlHIFa/VEGF axis in macrophagesWenext focused on the specific isoforms of PI3K involved

in regulating the HIFa/VEGF axis in macrophages. Recent-ly, published reports (26) and our results provide evidencethat p110g is the predominant isoform expressed in themyeloid cells and macrophages (Supplementary Fig. S1A).To determine the specificity of the PI3K signaling in HIFaaccumulation isoform-specific PI3K inhibitors, namely,GDC-0941 (p110a-specific inhibitor; ref. 36), TGX-221(p110b-specific inhibitor; ref. 37), Cal101 (p110d-specificinhibitor; ref. 38) and AS605240 (p110g-specific inhibitor;ref. 39) were used. Interestingly, we found that AS605240 (ap110g isoform inhibitor) showed the most marked inhibi-tion of HIFa accumulation and decreased the transcrip-tional activation of VEGF in BMDMs, whereas the effect ofinhibitors of a, b, and d isoform did not show a markedeffect on HIFa accumulation (Fig. 3A and B). However,inhibitors ofa,b, and d isoform significantly blockedVEGFmRNA expression. Although isoform-specific inhibitors ofPI3K display some level of selectivity, the kinase inhibitoryprofiles and IC50 potencies against different p110 isoformsare not absolute, making conclusions difficult. To confirm aspecific role for different PI3K catalytic isoforms in theregulation of HIF1a accumulation, we therefore knockeddown the expression of p110 isoforms in BMDMs by usingsiRNA against these isoforms. As shown in Fig. 3C and D,targeting of specific p110g isoform led to decreased hypoxia-induced HIFa stabilization and reduced expression ofVEGF mRNA, whereas there is no effect on HIFa proteinand VEGF mRNA levels on targeting a and b isoforms.Interestingly, knockdown of p110d isoform also blockedaccumulation of HIFa protein as well as VEGF geneexpression (Fig. 3C and D). Taken together, these resultssupport that HIFa stability is mainly contributed by p110gisoform and partially by p110d isoform of PI3K. To validateour results that p110g is the major isoform involved inregulating HIFa stability, we used p110g�/� mice. TheBMDMs isolated from these mice show decreased HIFaaccumulation and VEGF secretion (Fig. 3E and F). How-ever, p110g�/� BMDMs treated with AS605240 (p110g-

specific inhibitor) did not show any further reduction inVEGF levels (Fig. 3F), further validating our results thatp110g is the major isoform involved in HIFa accumulationand VEGF secretion.

Blocking PI3K signaling axis induces the hypoxicdegradation of HIFa via the 26S proteasome pathwayThe above results (Figs. 1–3) clearly establish a key role

for the PTEN/PI3K signaling pathway in hypoxia-induced HIFa accumulation in macrophages. To explorethe mechanism by which PTEN/PI3K pathway affectsHIFa levels in macrophages, we used the proteasomeinhibitor, MG132 to ask whether this process was relatedto proteasome-dependent degradation of HIF. BMDMsfrom PTENMC-WT and PTENMC-KO mice were treatedwith the proteasome inhibitor, MG132, before exposureto hypoxia. Our data support the hypothesis that PTENaffects HIFa under hypoxic conditions by inducing thehypoxic degradation of HIFa as this effect can be reversedby pretreatment with the proteasome inhibitor, MG132(Fig. 4A). To further confirm that the disappearance ofHIFa is a proteasome-dependent event and that PI3Kactivation is required for the hypoxic stability of HIFa, wetreated macrophages with a panel of PI3K inhibitors. Thedecrease in HIFa accumulation observed in BMDMstreated with pan-PI3K inhibitors, p110g isoform-specificPI3K inhibitor, AS605240, and novel pan-PI3K inhibi-tor, SF2523, was reversed by treatment with MG132 (Fig.4B–D). The degradation of HIFa by VHL under nor-moxic conditions is well known; however, its degradationunder hypoxia is poorly understood. Importantly, ourresults provide direct evidence that HIFa is degradedunder hypoxic conditions in a PTEN- and PI3K-depen-dent manner in macrophages.

PI3Kg signaling in macrophages promotes tumorgrowth and metastasisIt is well documented that macrophages exposed to

hypoxia accumulate both HIF1a and HIF2a, and over-expression of HIF2a in TAMs is specifically correlated withhigh-grade human tumors and poor prognosis (9). On thesame note, Doedens and colleagues reported that macro-phage expression of HIF1a suppresses T-cell function andpromotes tumor growth (40). On the basis of the resultsobserved in the present study showing the important role ofPI3K signaling axis in controlling both HIF1a and HIF2aaccumulation in macrophages, we hypothesize that PI3Ksignaling should promote tumor growth by increasing secre-tion of proangiogenic factors in TAMs. To validate thishypothesis, syngeneic LLCwas injected subcutaneously intoWT and p110g�/� mice. Tumor growth and angiogenesiswere drastically reduced in p110g�/� mice (Fig. 5A and B).ImmunofluorescenceCD31 staining showed that themicro-vessel density was significantly reduced in tumors propagat-ed in p110g�/� mice (Fig. 5B). On the basis of previouslypublished report that p110g is the isoform predominantlyexpressed in myeloid cell compartment (26) and our obser-vation that p110g isoform is the major contributor of

Joshi et al.





hypoxia-induced both HIF1a and HIF2a accumulation inmacrophages (Fig. 3C–F), we hypothesized that macro-phages infiltrating into the tumors implanted in p110g�/

� mice would be defective in secretion of proangiogenicfactors that promote tumor growth. To validate this hypoth-esis, macrophages were FACS sorted (on the basis of CD11band F4/80 staining) fromLLC tumors implanted inWT andp110g�/� mice and evaluated for expression of genesrequired for tumor growth and progression. For this, a panelof TAM-associated genes, VEGF, matrix metalloproteinase(MMP)9, urokinase-type plasminogen activator (uPA),COX2, arginase, MMR, TGFb, IL1, and TNFa, was usedto examine the gene expression profile of macrophagespresent in the tumor microenvironment of WT andp110g�/� mice. The expression levels of COX2, uPA,MMP9, MMR, arginase, and VEGF are significantly higherin macrophages sorted from LLC tumors of WT than in

p110g�/� mice (Fig. 5C). In contrast, the macrophagessorted from tumors implanted in p110g�/� animals showedhigher levels of IL1 and TNFa (Fig. 5C). Consistent withthis, existing literature suggests that TAMs secrete severalproteins includingMMP-9 (41), uPA (42), and VEGF (43),which promote tumor growth, angiogenesis, and metastasis.IL1 and TNFa are proinflammatory cytokines which arereported to be expressed at a lower level in tumor-associatedM2 macrophages and are known to promote tumor sup-pression (44). These results suggest that macrophages iso-lated from tumors implanted in p110g�/� mice expresshigher levels of proinflammatory cytokines seen in M1macrophages which promote tumor suppression, whereasmacrophages isolated fromWTmice show higher expressionof genes which promote tumor growth and progression.Tumor growth and angiogenesis are prerequisites for metas-tasis; hence, we examined whether p110g�/� mice display

Co

ntr

ol

100

nm

ol/L

GD

C09

41

100

nm

ol/L

TG

X22

1

100

nm

ol/L

Cal

101

100

nm

ol/L

AS

6052

40

HIF1a

Rel

ativ

e V

EG

F e

xpre

ssio

n

HIF2apAKT

AKT

WT

191 192 118 119 Mouse ID

HIF1a

HIF2a

b-Actin

p110g–/–

b-Actin

50 Ctrl

A

D E F

B CsiRNA

HIF1a

HIF2a

p110ap110b

p110gpAKTAKTb-Actin

p110d

a b d g

40

30

20

10

0

Rel

ativ

e V

EG

F e

xpre

ssio

n

50

40

30

20

10

0

Rel

ativ

e V

EG

F e

xpre

ssio

n

50

40

30

20

10

0

Control

siRNA p11

0a

siRNA

p110b

siRNA

p110d

siRNA

p110g

siRNA

Control

100 n

mol/L

GDC09

41

100 n

mol/L

TGX221

100 n

mol/L

Cal1

01

100 n

mol/L

AS60

5240

WT p110g–/– p110g–/–

+ 100 nmol/L AS605240

Figure 3. Specific p110g andp110d isoformof PI3Kcontrol HIFa/VEGF axis inmacrophages. A andB,BMDM fromWTanimalswere treatedwith 100 nmol/L ofGDC0941or TGX221orCal101orAS605240 followedbyhypoxia for 4 hours forWestern blot analysis or 24hours forRNA isolation and real-timePCRanalysis.C and D, BMDMs fromWT animals were transfected with either 100 nmol/L control siRNA or siRNA against p110 a or b or d or g isoform using AmaxaMouseMacrophage Nucleofaction kit (Lonza). Thirty-six hours after transfection, hypoxia was given to these siRNA-transfected BMDMs for 4 hours forWestern blotanalysis (C) or 24 hours for RNA isolation and real-time PCR analysis (D). E, Western blot analysis of HIF1a and HIF2a from WT and p110g�/� BMDMskept in hypoxic (1% O2) conditions for 4 hours followed by preparation of nuclear extracts. Each lane in figure represents lysate prepared from an individualmouse. F, relative VEGFexpression inWT, p110 g�/�BMDMs treatedwith orwithout 100 nmol/LAS605240 and kept in hypoxia (1%O2) for 24 hours, followedby RNA isolation and real-time PCR studies. Graphs in B, D, and F represent mean� SEM (n¼ 4–5) for 3 independent experiments. Statistical significance isassessed by 2-sample t test. �, P < 0.05; ��, P < 0.01; ��, P < 0.001.






any defect in metastasis. To explore the role of p110g inexperimental lung metastasis, B16F10 melanoma cells wereinjected into mouse tail veins ofWT and p110g�/� animals.Two weeks later, we observedmassive numbers of metastaticfoci in the lungs of WT mice compared with a markedreduction in number of nodules in p110g�/� mice(Fig. 5D). A microscopic analysis of lung tissue sectionsconfirmed the results from gross specimens and demonstrat-ed larger numbers of metastatic foci in the lungs ofWTmice(Fig. 5D). These observations led us to examine whether theBMDMs prepared from p110g�/�mice in vitro display anydefect in expression of genes promoting tumor growth andprogression. We found that macrophages isolated from LLCtumors growing in p110g�/� mice were defective in theexpression of genes previously implicated in the promotionof tumor growth (M2 markers; Fig. 5E) including VEGFdescribed in Fig. 3F.

SF1126, a pan-PI3K inhibitor, blocks tumor growth andmetastasis and reduces the expression of proangiogenicfactors by TAMsConsidering the role of PTEN and PI3K signaling in

the macrophage-dependent promotion of tumor growthand tumor-induced angiogenesis, we examined the effectsof a pan-PI3K inhibitor, SF1126, in the LLC model for itseffects on tumor growth, angiogenesis, and metastasis invivo (Fig. 6). SF1126 has completed phase I clinical trial insolid tumors and B-cell malignancies and is headed for aphase II trial in 2014 (29). The administration of 50

mg/kg dose of SF1126, 3 times a week, significantlyblocked tumor growth in WT animals implanted withLLC (Fig. 6A). Hence, we determined whether SF1126exerts an effect on the neovascularization of these tumors.Immunofluoresence CD31 staining showed that themicrovessel density was significantly reduced in the micetreated with SF1126 (Fig. 6B). Because TAMs are knownto have a marked influence on the regulation of tumor-induced angiogenesis (45), we propose that SF1126 blocksangiogenesis in the tumors by blocking the proangiogenicfactors secreted by TAMs within the tumor microenvi-ronment. The supporting evidence for this notion comesfrom the real-time PCR data obtained from macrophages(FACS sorted on basis of CD11b and F4/80 staining)isolated from the LLC tumors implanted into WT miceand treated with or without SF1126 (Fig. 6C). Interest-ingly, the macrophages sorted from the mice treated withSF1126 inhibitor showed reduced expression of VEGF,MMP9, arginase, and COX2 (Fig. 6C). The results in Fig.6C suggest that macrophages isolated from SF1126-trea-ted tumors display a reduced expression of proangiogenicfactors promoting metastasis. So, we next explored wheth-er the administration of SF1126 can block tumor metas-tasis. We observed a marked increase in metastatic foci inthe lungs of WT mice compared with the 60% reductionin metastatic nodules seen in SF1126-treated miceinjected with B16F10 melanoma (P < 0.01; Fig. 6D andE). Our results demonstrating that SF1126 blockedmetastasis lead us to explore the effect of SF1126 on

PT

EN

MC

-WT

PT

EN

MC

-ko

PT

EN

MC

-WT

PT

EN

MC

-ko

Co

ntr

ol

Co

ntr

ol

Co

ntr

ol

Co

ntr

ol

25 m

mo

l/L S

F11

26

10 m

mo

l/L S

F25

23

100

nm

ol/L

AS

6052

40

Co

ntr

ol

100

nm

ol/L

AS

6052

40

Co

ntr

ol

100

nm

ol/L

AS

6052

40

Co

ntr

ol

100

nm

ol/L

AS

6052

40

10 m

mo

l/L S

F25

23

500

nm

ol/L

PF

4691

502

Co

ntr

ol

25 m

mo

l/L S

F11

26

500

nm

ol/L

PF

4691

502

Co

ntr

ol

25 m

mo

l/L S

F11

26

500

nm

ol/L

PF

4691

502

Co

ntr

ol

25 m

mo

l/L S

F11

26

500

nm

ol/L

PF

4691

502

MG132

A B

C D

– – + +

HIF1aHIF2a

b-Actin

Normoxia Hypoxia– – – –+ + + + + + MG132

HIF1aHIF2ab-Actin

MG132– – + +

HIF1a

HIF2ab-Actin

MG132– – – – + +++

Normoxia Hypoxia

HIF1aHIF2ab-Actin

– –

Figure 4. Hypoxic degradation ofHIF1a and HIF2a by PTEN/PI3K/AKT pathway occurs via the 26Sproteasome. A, Western blotanalysis of HIF1a and HIF2a fromBMDMs from PTENMC-WT andPTENMC-KO treated with 10 mmol/L MG132 and kept in hypoxicconditions for 4 hours. B–D, WTBMDMswere treatedwith 10mmol/Lproteasome inhibitorMG132 for 5minutes before being pulsed withthe pan-PI3K inhibitors, 25 mmol/LSF1126, and 500 nmol/LPF4691502 (B), 100 nmol/LAS605240 (C), and 10 mmol/LSF2523 (D) followed by normoxia(21% O2) or hypoxia (1% O2) for4 hours. Nuclear extracts wereprepared for HIFa. Western blotanalyses are as described before.Experiments were repeated 3 to4 times with similar results.

Joshi et al.





survival in the B16 experimental metastasis model. Thedata in Fig. 6F demonstrate that SF1126 treatment pro-longs the survival of mice by average 15 days, a resultwhich directly correlates with the inhibition of experi-mental metastasis in this model and validating the efficacyof this drug either alone or in combination with otherdrugs in the treatment of cancer. Taken together, weconclude that SF1126 blocks tumor growth and metastasispotentially by blocking HIF1a and HIF2a and thedownstream expression and secretion of proangiogenicfactors by tumor-infiltrating M2 macrophages. On thebasis of these results, we propose a hypothetical model inwhich E3 ligase degrades HIF1a under hypoxic condi-tions in the cytoplasm in a proteasome-dependent man-ner. The model would predict that the stimulation of

PI3K and AKT via upstream signals will serve to promotetranslocation of E3 ligase into the nucleus and hencepromote the stabilization of hypoxic HIF1a. The pathwaywill bring hypoxic HIF1a under the control of the PTEN/PI3K/AKT signaling axis to regulate an E3 ligase tocontrol hypoxic HIF1a levels and angiogenesis and tumormetastasis in vivo (Fig. 7).

DiscussionIt is well established that TAMs accumulate in hypoxic

regions of tumors and that hypoxia triggers a proangiogenicprogram in these cells stimulating the production of angio-genic factors and matrix-regulatory proteins such as VEGFand MMPs, respectively (46, 47). Therefore, macrophages

2,500

2,000

1,500

1,000

11 14 17Days

WT

WT BMDM

p110g–/–

p110g–/– BMDM

p110g–/–p110g–/–

WT p110g–/– 25.0

20.0

15.0

10.0

5.0

10.0

7.5

5.0

Rel

ativ

e g

ene

exp

ress

ion

2.5

0

0WT p110g–/–

CD

31

Sta

inin

g

Mic

rove

ssel

No

./HP

F

Tum

or

volu

me

(mm

3 )

21 25

IL1

TNFauPA

MMP9MMR

TGFb

500

0

175

150

125

100

75

50

25

0Rel

ativ

e g

ene

exp

ress

ion

IL1

Cox2

uPAMMR

MM

P 9

VEGF

Argin

ase

TNFa

WT

WT

75

50

25

0

Nu

mb

er o

f m

etas

tati

c n

od

ule

s

A B

CE

D

Macrophages sorted from LLC tumors of WT miceMacrophages sorted from LLC tumors of p110g–/– mice

Figure 5. Expression of p110g isoform inmacrophages promotes tumor growth and angiogenesis. A, tumor growth of LLC cells inoculated subcutaneously inWT and p110g�/�mice (n¼ 6–8). B, left, representative immunofluorescent staining of tumor vasculature by CD31 (green) and counterstain by DAPI (blue) onfrozen tumor sectionsof LLC tumors implanted subcutaneously inWTandp110g�/�mice.Right, reducedmicrovascular density (MVD) in tumors isolated fromp110g�/� animals as compared withWT animals. MVDwas determined by counting the number ofmicrovessels per high-power field (HPF) in the section withan antibody reactive to CD31.Microvessels were counted blindly in 5 to 10 randomly chosen fields, and data are representative of 3 independent experimentswith 4 to 5 mice. �, P < 0.05 versus WT. C, relative expression levels of specific genes required for tumor growth in macrophages sorted from LLCtumors implanted intoWTandp110g�/� animals. D, left, representative photograph of pulmonarymetastatic foci produced 15 days after intravenous injectionof B16F10 cells in WT and p110g�/� mice. Right, number of experimental pulmonary B16F10 metastases in lungs counted under dissecting microscope.E, expression of specific tumor-promoting genes in the BMDMs isolated from WT and p110g�/� mice and cultured in MCSF in vitro. Graphs presentmean � SEM of 7 mice for A, B, and D and 3 to 4 mice for C. �, P < 0.05; ��, P < 0.01; ��, P < 0.001 versus WT, as determined using 2-sample t test.






recruited in situ represent an indirect pathway of amplifi-cation of angiogenesis, in concert with angiogenic moleculesdirectly produced by tumor cells. Thus, we speculate thattargeting the common pathway promoting angiogenesis inmacrophages as well as tumor cell compartment will likelyprovide more effective therapy in the treatment of cancer.The PI3K pathway is known to play a critical role in tumorcell survival, angiogenesis, invasion, apoptosis, and cellularmetabolism (15, 22) and is considered as themost frequentlyactivated pathway in cancer. Herein, we provide directevidence that PI3K pathway controls tumor growth and

angiogenesis and regulates the HIFa/VEGF axis in macro-phages. Moreover, SF1126, a pan-PI3K inhibitor enteringphase II clinical trials, blocks macrophage-mediated tumorgrowth, angiogenesis, and metastasis, suggesting efficacy ofSF1126 and other PI3K inhibitors in the treatment of cancerin which macrophages are important in tumor progression.Recent studies have clearly demonstrated that HIF1a and

HIF2a are stabilized in hypoxic macrophages in vitro (12).Talks and colleagues demonstrated that HIF2a proteinaccumulates at high levels in TAMs detected in varioushuman cancers (9). Recent reports suggest the important role

3,5003,0002,5002,0001,5001,000

5000

0 3 6 9 12 15 18

Time of treatment (days)

Vehicle

Vehicle

Vehicle

SF1126 treated

SF1126 treated

SF1126 treated

SF1126 treated

SF1126 treated

Tum

or

volu

me

(mm

3 )30

20

10

0

Mic

rove

ssel

s N

o./H

PF

CD

31 S

tain

ing

250

200

150

100

50

0

IL1

IL6

Cox2

MMP9

MMR

Argin

ase

VEGF

Rel

ativ

e g

ene

exp

ress

ion

Vehicle SF1126treated

Vehicle

Vehicle

Vehicle

SF1126treated

60

50

40

30

20

10

0

Nu

mb

er o

f m

etas

tati

c n

od

ule

s

Days

Per

cen

t su

rviv

al

100

80

60

40

20

00 5 10 15 20 25 30 35 40

Macrophages isolated from LLC tumors of untreated miceMacrophages isolated from LLC tumors of mice treated with SF1126

A B

C E

D

F

Figure6. SF1126, a pan-PI3K inhibitor blocks tumor growth, angiogenesis,metastasis and reduces the expression of proangiogenic factors byTAMs. A, tumorvolume of LLC inoculated subcutaneously in WT mice treated with or without 50 mg/kg SF1126 3 times a week. Treatment was started 10 days afterinoculation of LLC cells and was continued until tumors were harvested (n ¼ 6–8). B, MVD was scored by manual counting of CD31-positive vessels insubcutaneous LLC tumor sections was reduced in tumors isolated from WT animals treated with SF1126. Right, reduced microvascular density in tumorsisolated fromuntreated or SF1126-treated animals as confirmedby quantification ofCD31-positive area in high-power field (HPF) C, relative expression levelsof specific genes required for tumor growth in macrophages sorted from LLC tumors implanted intoWTmice and treated with 50mg/kg of SF1126, 3 times aweek until tumors were harvested. D, left, representative photograph of pulmonary metastatic foci produced in WT animals injected with B16F10 melanomaand treated with daily dose of 50 mg/kg SF1126 for 15 days. Right, number of experimental pulmonary B16F10 metastases in lungs. E, figure showsrepresentative lung hematoxylin and eosin (H&E) sections of B16 metastases from WT mice treated with or without 50 mg/kg SF1126. Arrows point tometastatic foci. F, Kaplan–Meier survival curve for B16 metastasized mice treated with or without 50 mg/kg SF1126 daily. Data are representative of 2 to 3independent experiments. Graphs present mean � SEM of 8 to 10 mice for A, B, and D–F and 3 to 4 mice for C. �, P < 0.05 and ��, P < 0.01 versus WT, asdetermined using 2-sample t test.

Joshi et al.





of HIF2a in macrophage functions in mouse models oftumor inflammation (48). In the present study, using geneticmouse models, siRNA approach, and clinically relevantPI3K inhibitors, we provide evidence that PTEN/PI3Ksignaling controls both HIF1a and HIF2a accumulationin macrophages. The role of PTEN/PI3K signaling inpromoting HIFa stabilization in tumor cells is well-docu-mented. However, there are no reports examining the role ofthis important signaling axis inHIFa stabilization in macro-phages. Our studies clearly demonstrate that the loss ofPTEN in the myeloid cell compartment promotes hypoxicHIFa stabilization in BMDMs (Fig. 1). Our studies provideconvincing evidence that the expression of PTEN andinhibition of PI3K/AKT signaling induces the hypoxicdegradation of HIF1a and HIF2a in a proteasome-depen-dent manner. While much is known about the transcrip-tional and posttranslational regulation of HIFa, there is verylimited evidence that HIFa can be degraded under condi-tions of hypoxia by the 26S proteasome (49). Liu andcolleagues recently described oxygen-independent degrada-tion of HIF1a by a novel HIF1a-interacting protein,receptor for activated C-kinase 1 (RACK1; ref. 50). Theysuggested that similar to the E3 ligase pVHL, RACK1increased polyubiquitination of HIF1a and was unable tomediate degradation of HIF1a in the presence of protea-

somal inhibitor MG132. In the same context, recent studysuggests that SHARP promotes proteasomal degradation ofHIF1a independent of pVHL, hypoxia, and ubiquitinationmachinery (51). Recently published work from our labora-tory performed in glioma tumor models has demonstratedthat PTEN and PI3K inhibitors control the hypoxic deg-radation of HIF1a and that the E3 ligase, MDM2, isrequired for hypoxic degradation of HIF1a (52) The E3ligase and other signaling pathways controlling the degra-dation of HIF1a and HIF2a under hypoxic conditions inmacrophages remain an active area of investigation and willbe a major focus of our future research efforts.Moreover, we observed that p110g , the most dominant

isoform of PI3K expressed in macrophages, controls HIF1aand HIF2a accumulation in macrophages. Interestingly,our data provide evidence that p110d isoform contributessignificantly to HIFa stabilization and VEGF secretion inBMDMs (Fig. 3C and D). More importantly, we demon-strate that the expression of p110g inmacrophages promotestumor growth and metastasis potentially by controlling theexpression of proangiogenic and tumor growth-promotingfactors by TAMs (Fig. 5C and E). A recent report by Schmidand colleagues has shown that p110g promotes tumorinflammation by controlling the trafficking of myeloid cellsinto tumor (26). Our work extends these findings and

Chemokines

Macrophage

Metastasis

Tumor progression

PI3Kg

PI3K inhibitors

E3 ligase E3 ligase

Epo

Glycolyticenzymes

Nos2

VEGF

P

P

AKTPTEN

PIP2

PIP3

Hypoxia

HIFa

HIFa

HIFb

HIFa

HIFa

HIFb HRE

p300

Transcription

Nucleus

GPCR

Proteasomaldegradation

Ubiquitination

of HIFa

Figure 7. Mechanism by which the PTEN/PI3K/AKT signaling axis exerts control over HIFa stability, tumor growth, and metastasis by secretion ofproangiogenic factors by macrophages. PI3K inhibitors block the phosphorylation of an E3 ligase resulting in cytoplasmic localization of E3 ligase anddegradation of HIF1a; this could have therapeutic implications for HIF1a/VEGF axis in PI3K inhibitor cancer therapeutics.






suggests that the expression of p110g in macrophages con-trols the hypoxia-driven HIFa/VEGF axis which plays acrucial role in neovascularization and tumor growth. More-over, our studies provide evidence that administration ofSF1126 blocks macrophage-mediated tumor growth andangiogenesis. SF1126 has completed phase I clinical trial inB-cell malignancies and solid tumors with excellent safetyprofile (29). It is currently slated for a phase II trial inPIK3CA-mutated human cancers. Herein, we demonstratethat SF1126 which potently blocks tumor proliferation andsurvival (29, 53) exerts a marked effect on the tumor stromalmacrophage compartment as it blocks tumor growth, angio-genesis, and metastasis in vivo. In conclusion, the majorfindings of this article are: (i) we identified an important roleof PTEN/PI3K/AKT signaling axis in HIF1a and HIF2astabilization inmacrophages; (ii) PTENand several clinicallyrelevant PI3K inhibitors exert their control over HIFa/VEGF axis by inducing the hypoxic degradation of HIFasubunit in a proteasome-dependent manner; (iii) expressionof p110g in macrophages promotes tumor growth andmetastasis; and (iv) SF1126 potently blocked macro-phage-mediated tumor angiogenesis and metastasis. Thesestudies suggest that SF1126 and other PI3K inhibitors areexcellent candidates for the treatment of cancers wheremacrophages promote tumor progression.

Disclosure of Potential Conflicts of InterestD.L. Durden discloses a financial conflict of interest relating to the development

of SF1126. D.L. Durden has ownership interest (including patents) in SignalRxPharmaceuticals. This relationship has been reviewed by the UCSD committee onconflict of interest. No potential conflicts of interest were disclosed by the otherauthors.

Authors' ContributionsConception and design: S. Joshi, A.R. Singh, D.L. DurdenDevelopment of methodology: S. Joshi, A.R. Singh, D.L. DurdenAcquisition of data (provided animals, acquired and managed patients, providedfacilities, etc.): M. ZulcicAnalysis and interpretation of data (e.g., statistical analysis, biostatistics, compu-tational analysis): S. Joshi, D.L. DurdenWriting, review, and/or revision of the manuscript: S. Joshi, A.R. Singh,D.L. DurdenAdministrative, technical, or material support (i.e., reporting or organizing data,constructing databases): D.L. DurdenStudy supervision: D.L. Durden

Grant SupportThis research effort was funded by CA94233 and HL091385 from NIH to D.L.

Durden and grants from ALSF (Springboard Grant) and Hyundai Hope on WheelsFoundation, Hope Grant.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be herebymarked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

Received January 21, 2014; revised July 21, 2014; accepted July 27, 2014;published OnlineFirst August 7, 2014.

References1. Mantovani A, Schioppa T, Porta C, Allavena P, Sica A. Role of tumor-

associated macrophages in tumor progression and invasion. CancerMetastasis Rev 2006;25:315–22.

2. Mantovani A, Sica A. Macrophages, innate immunity and cancer:balance, tolerance, and diversity. Curr Opin Immunol 2010;22:231–7.

3. Wang GL, Semenza GL. General involvement of hypoxia-induciblefactor 1 in transcriptional response to hypoxia. Proc Natl Acad SciU S A 1993;90:4304–8.

4. Weidemann A, Johnson RS. Biology of HIF-1alpha. Cell Death Differ2008;15:621–7.

5. Semenza GL. Hypoxia-inducible factor 1: oxygen homeostasis anddisease pathophysiology. Trends Mol Med 2001;7:345–50.

6. Kallio PJ, Wilson WJ, O'Brien S, Makino Y, Poellinger L. Regulation ofthe hypoxia-inducible transcription factor 1alpha by the ubiquitin-proteasome pathway. J Biol Chem 1999;274:6519–25.

7. Cramer T, Yamanishi Y, Clausen BE, Forster I, Pawlinski R, MackmanN, et al. HIF-1alpha is essential for myeloid cell-mediated inflamma-tion. Cell 2003;112:645–57.

8. Semenza G. Signal transduction to hypoxia-inducible factor 1. Bio-chem Pharmacol 2002;64:993–8.

9. Talks KL, Turley H, Gatter KC, Maxwell PH, Pugh CW, Ratcliffe PJ,et al. The expression and distribution of the hypoxia-induciblefactors HIF-1alpha and HIF-2alpha in normal human tissues, can-cers, and tumor-associated macrophages. Am J Pathol 2000;157:411–21.

10. Leek RD, Talks KL, Pezzella F, Turley H, Campo L, Brown NS, et al.Relation of hypoxia-inducible factor-2 alpha (HIF-2 alpha) expressionin tumor-infiltrative macrophages to tumor angiogenesis and theoxidative thymidine phosphorylase pathway in Human breast cancer.Cancer Res 2002;62:1326–9.

11. White JR, Harris RA, Lee SR, Craigon MH, Binley K, Price T, et al.Genetic amplification of the transcriptional response to hypoxia as anovel means of identifying regulators of angiogenesis. Genomics2004;83:1–8.

12. Burke B, Tang N, Corke KP, Tazzyman D, Ameri K, Wells M, et al.ExpressionofHIF-1alpha byhumanmacrophages: implications for the

use of macrophages in hypoxia-regulated cancer gene therapy.J Pathol 2002;196:204–12.

13. Bardos JI, Ashcroft M. Hypoxia-inducible factor-1 and oncogenicsignalling. Bioessays 2004;26:262–9.

14. Hirota K, Semenza GL. Rac1 activity is required for the activation ofhypoxia-inducible factor 1. J Biol Chem 2001;276:21166–72.

15. Jiang BH, Jiang G, Zheng JZ, Lu Z, Hunter T, Vogt PK. Phosphatidy-linositol 3-kinase signaling controls levels of hypoxia-inducible factor1. Cell Growth Differ 2001;12:363–9.

16. Zhong H, Chiles K, Feldser D, Laughner E, Hanrahan C, GeorgescuMM, et al. Modulation of hypoxia-inducible factor 1alpha expressionby the epidermal growth factor/phosphatidylinositol 3-kinase/PTEN/AKT/FRAP pathway in human prostate cancer cells: implications fortumor angiogenesis and therapeutics. Cancer Res 2000;60:1541–5.

17. ZundelW,SchindlerC,Haas-KoganD,KoongA,Kaper F,ChenE, et al.Loss of PTEN facilitates HIF-1-mediated gene expression. Genes Dev2000;14:391–6.

18. Alvarez-Tejado M, Alfranca A, Aragones J, Vara A, Landazuri MO, delPeso L. Lack of evidence for the involvement of the phosphoinositide3-kinase/Akt pathway in the activation of hypoxia-inducible factors bylow oxygen tension. J Biol Chem 2002;277:13508–17.

19. Arsham AM, Plas DR, Thompson CB, SimonMC. Phosphatidylinositol3-kinase/Akt signaling is neither required for hypoxic stabilization ofHIF-1 alpha nor sufficient for HIF-1-dependent target gene transcrip-tion. J Biol Chem 2002;277:15162–70.

20. Vanhaesebroeck B, Leevers SJ, Ahmadi K, Timms J, Katso R, DriscollPC, et al. Synthesis and function of 3-phosphorylated inositol lipids.Annu Rev Biochem 2001;70:535–602.

21. Cantley LC. The phosphoinositide 3-kinase pathway. Science 2002;296:1655–7.

22. Castellino RC, Muh CR, Durden DL. PI-3 kinase-PTEN signaling node:an intercept point for the control of angiogenesis. Curr Pharm Des2009;15:380–8.

23. Munugalavadla V, Borneo J, Ingram DA, Kapur R. p85alpha subunit ofclass IA PI-3 kinase is crucial for macrophage growth and migration.Blood 2005;106:103–9.

Joshi et al.





24. Curnock AP, Logan MK, Ward SG. Chemokine signalling: pivotingaround multiple phosphoinositide 3-kinases. Immunology 2002;105:125–36.

25. Hirsch E, Katanaev VL, Garlanda C, Azzolino O, Pirola L, Silengo L,et al. Central role for G protein-coupled phosphoinositide 3-kinasegamma in inflammation. Science 2000;287:1049–53.

26. Schmid MC, Avraamides CJ, Dippold HC, Franco I, Foubert P, ElliesLG, et al. Receptor tyrosine kinases and TLR/IL1Rs unexpectedlyactivatemyeloid cell PI3kgamma, a single convergent point promotingtumor inflammation and progression. Cancer Cell 2011;19:715–27.

27. Schabbauer G, Matt U, Gunzl P, Warszawska J, Furtner T, Hainzl E,et al. Myeloid PTEN promotes inflammation but impairs bactericidalactivities during murine pneumococcal pneumonia. J Immunol2010;185:468–76.

28. Li Y, Jia Y, Pichavant M, Loison F, Sarraj B, Kasorn A, et al. Targeteddeletion of tumor suppressor PTEN augments neutrophil function andenhances host defense in neutropenia-associated pneumonia. Blood2009;113:4930–41.

29. Mahadevan D, Chiorean EG, Harris WB, VonHoff DD, Stejskal-BarnettA,QiW, et al. Phase I pharmacokinetic andpharmacodynamic study ofthe pan-PI3K/mTORC vascular targeted pro-drug SF1126 in patientswith advanced solid tumours and B-cell malignancies. Eur J Cancer2012;48:3319–27.

30. Morales GA, Garlich JR, Su J, Peng X, Newblom J, Weber K, et al.Synthesis and cancer stem cell-based activity of substituted 5-mor-pholino-7H-thieno[3,2-b]pyran-7-ones designed as next generationPI3K inhibitors. J Med Chem 2013;56:1922–39.

31. Joshi S, Singh AR, Zulcic M, Bao L, Messer K, Ideker T, et al. Rac2controls tumor growth, metastasis and m1-m2 macrophage differen-tiation in vivo. PLoS ONE 2014;9:e95893.

32. AndrewsNC, Faller DV. A rapidmicropreparation technique for extrac-tion of DNA-binding proteins from limiting numbers of mammaliancells. Nucleic Acids Res 1991;19:2499.

33. Yuan J,Mehta PP, YinMJ, Sun S, Zou A, Chen J, et al. PF-04691502, apotent and selective oral inhibitor of PI3K and mTOR kinases withantitumor activity. Mol Cancer Ther 2011;10:2189–99.

34. Bendell JC, Rodon J, Burris HA, de Jonge M, Verweij J, Birle D, et al.Phase I, dose-escalation study of BKM120, an oral pan-Class I PI3Kinhibitor, in patients with advanced solid tumors. J Clin Oncol 2012;30:282–90.

35. FanQW,Knight ZA,GoldenbergDD, YuW,MostovKE, StokoeD, et al.A dual PI3 kinase/mTOR inhibitor reveals emergent efficacy in glioma.Cancer Cell 2006;9:341–9.

36. Folkes AJ, Ahmadi K, Alderton WK, Alix S, Baker SJ, Box G, et al. Theidentification of 2-(1H-indazol-4-yl)-6-(4-methanesulfonyl-piperazin-1-ylmethyl)-4-morpholin-4-yl-t hieno[3,2-d]pyrimidine (GDC-0941) asa potent, selective, orally bioavailable inhibitor of class I PI3 kinase forthe treatment of cancer. J Med Chem 2008;51:5522–32.

37. Jackson SP, Schoenwaelder SM, Goncalves I, Nesbitt WS, Yap CL,Wright CE, et al. PI 3-kinase p110beta: a new target for antithrombotictherapy. Nat Med 2005;11:507–14.

38. Lannutti BJ, Meadows SA, Herman SE, Kashishian A, Steiner B,Johnson AJ, et al. CAL-101, a p110delta selective phosphatidylino-sitol-3-kinase inhibitor for the treatment of B-cellmalignancies, inhibitsPI3K signaling and cellular viability. Blood 2011;117:591–4.

39. Wei X, Han J, Chen ZZ, Qi BW, Wang GC, Ma YH, et al. A phosphoi-nositide 3-kinase-gamma inhibitor, AS605240 prevents bleomycin-induced pulmonary fibrosis in rats. Biochem Biophys Res Commun2010;397:311–7.

40. DoedensAL, StockmannC, RubinsteinMP, Liao D, ZhangN, DeNardoDG, et al. Macrophage expression of hypoxia-inducible factor-1 alphasuppresses T-cell function and promotes tumor progression. CancerRes 2010;70:7465–75.

41. Giraudo E, Inoue M, Hanahan D. An amino-bisphosphonate targetsMMP-9-expressing macrophages and angiogenesis to impair cervicalcarcinogenesis. J Clin Invest 2004;114:623–33.

42. Marconi C, Bianchini F, Mannini A, Mugnai G, Ruggieri S, Calorini L.Tumoral and macrophage uPAR and MMP-9 contribute to the inva-siveness of B16 murine melanoma cells. Clin Exp Metastasis 2008;25:225–31.

43. Schoppmann SF, Fenzl A, Nagy K, Unger S, Bayer G, Geleff S, et al.VEGF-C expressing tumor-associated macrophages in lymph nodepositive breast cancer: impact on lymphangiogenesis and survival.Surgery 2006;139:839–46.

44. Sica A, Schioppa T, Mantovani A, Allavena P. Tumour-associatedmacrophagesare adistinctM2polarisedpopulationpromoting tumourprogression: potential targets of anti-cancer therapy. Eur J Cancer2006;42:717–27.

45. Knowles H, Leek R, Harris AL. Macrophage infiltration and angiogen-esis in human malignancy. Novartis Found Symp 2004;256:189–200;discussion 200–4, 259–69.

46. Lewis C, Murdoch C. Macrophage responses to hypoxia: implicationsfor tumor progression and anti-cancer therapies. Am J Pathol 2005;167:627–35.

47. Du R, Lu KV, Petritsch C, Liu P, Ganss R, Passegue E, et al. HIF1alphainduces the recruitment of bone marrow-derived vascular modulatorycells to regulate tumor angiogenesis and invasion. Cancer Cell 2008;13:206–20.

48. Imtiyaz HZ, Williams EP, Hickey MM, Patel SA, Durham AC, Yuan LJ,et al. Hypoxia-inducible factor 2alpha regulates macrophage functionin mouse models of acute and tumor inflammation. J Clin Invest2010;120:2699–714.

49. Berta MA, Mazure N, Hattab M, Pouyssegur J, Brahimi-Horn MC.SUMOylation of hypoxia-inducible factor-1alpha reduces its transcrip-tional activity. Biochem Biophys Res Commun 2007;360:646–52.

50. Liu YV, Baek JH, Zhang H, Diez R, Cole RN, Semenza GL. RACK1competes with HSP90 for binding to HIF-1alpha and is required for O(2)-independent and HSP90 inhibitor-induced degradation of HIF-1alpha. Mol Cell 2007;25:207–17.

51. Montagner M, Enzo E, Forcato M, Zanconato F, Parenti A, Ram-pazzo E, et al. SHARP1 suppresses breast cancer metastasis bypromoting degradation of hypoxia-inducible factors. Nature 2012;487:380–4.

52. Joshi S, Singh AR, Durden DL. MDM2 regulates hypoxic HIF1alphastability in an E3 ligase proteasome and PTEN-PI-3 kinase-AKTdependent manner. J Biol Chem 2014;289:22785–97.

53. Garlich JR, De P, Dey N, Su JD, Peng X, Miller A, et al. A vasculartargeted pan phosphoinositide 3-kinase inhibitor prodrug, SF1126,with antitumor and antiangiogenic activity. Cancer Res 2008;68:206–15.






2014;12:1520-1531. Published OnlineFirst August 7, 2014.Mol Cancer Res Shweta Joshi, Alok R. Singh, Muamera Zulcic, et al. Metastasis

Stability and Tumor Growth, Angiogenesis, andα and HIF2αHIF1A Macrophage-Dominant PI3K Isoform Controls Hypoxia-Induced

Updated version

10.1158/1541-7786.MCR-13-0682doi:

Access the most recent version of this article at:

Material

Supplementary

http://mcr.aacrjournals.org/content/suppl/2014/08/16/1541-7786.MCR-13-0682.DC1

Access the most recent supplemental material at:

Cited articles

http://mcr.aacrjournals.org/content/12/10/1520.full#ref-list-1

This article cites 53 articles, 20 of which you can access for free at:

Citing articles

http://mcr.aacrjournals.org/content/12/10/1520.full#related-urls

This article has been cited by 9 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://mcr.aacrjournals.org/content/12/10/1520To request permission to re-use all or part of this article, use this link



http://mcr.aacrjournals.org/lookup/doi/10.1158/1541-7786.MCR-13-0682

http://mcr.aacrjournals.org/content/suppl/2014/08/16/1541-7786.MCR-13-0682.DC1

http://mcr.aacrjournals.org/content/12/10/1520.full#ref-list-1

http://mcr.aacrjournals.org/content/12/10/1520.full#related-urls

http://mcr.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://mcr.aacrjournals.org/content/12/10/1520


Documents

A Macrophage-Dominant PI3K Isoform Controls …...Shweta Joshi 1, Alok R. Singh , Muamera Zulcic , and Donald L. Durden1,2,3 Abstract Tumor growth, progression, and response to the