EGFR- and VEGF(R)-Targeted Small Molecules ShowSynergistic Activity in Colorectal Cancer ModelsRefractory to Combinations of Monoclonal Antibodies
Virginie Poindessous1,2,3, Djamila Ouaret1,2,3, Karima El Ouadrani1,2,3, Aude Battistella1,2,3,Virginie F. M�egalophonos1,2,3, Nyam Kamsu-Kom1,2,3, Am�elie Petitprez1,2,3,Alexandre E. Escargueil1,2,3, Pascaline Boudou1,2,3, Sylvie Dumont3,4, Pascale Cervera3,4,Jean-Francois Fl�ejou2,3,4, Thierry Andr�e1,2,3,5, Christophe Tournigand1,2,3,5, Benoist Chibaudel5,Aimery de Gramont1,2,3,5, and Annette K. Larsen1,2,3
AbstractPurpose: Epidermal growth factor receptor (EGFR) and VEGF(R) signaling show extensive cross-talk,
providing a rationale for joint targeting of the two pathways. However, combinations of monoclonal
antibodies (mAb) targeting EGFR and VEGF showed disappointing activity in patients with colorectal
cancer (CRC). We speculated that inhibition of surface receptors and ligands might only partly prevent
oncogenic signaling whereas small-molecule tyrosine kinase inhibitors (TKI) would also influence
intracellular signaling.
Experimental Design: Mice with CRC xenografts were treated with two TKIs, vargatef and afatinib, or
with two mAbs, bevacizumab and cetuximab, and their influence on tumor growth, viability, in vivo DNA
synthesis, and the presence of phosphorylated EGFR and VEGFR was determined. The activity of the TKIs
was further characterized in CRC cells with different KRAS status.
Results:Vargatef and afatinib together showed strong tumor growth inhibition towardHT-29 xenografts
compared with either drug alone, which was associated with a 5-fold increase in apoptotic tumor cell death.
In comparison, bevacizumab and cetuximab together were exclusively cytostatic with nomore activity than
either drug alone. Exposure to the two TKIs was accompanied by a marked decrease of tumor-associated
intracellular phospho-VEGFR1 and phospho-EGFR, whereas similar exposure to the two mAbs had no
detectable effect. A synergistic activity of vargatef plus afatinib was observed in all eight CRC cell lines
examined, independent of KRAS status.
Conclusions: Our results indicate that attenuation of intracellular EGFR and/or VEGF signaling is
required for cytotoxic activity. These findings provide a rationale for trials of the TKIs, even in patients with
mutant KRAS. Clin Cancer Res; 17(20); 6522–30. �2011 AACR.
Introduction
The development of anticancer agents targetingoncogenic signaling pathways represents a major concep-tual breakthrough. However, in many cases, the clinical
outcome has been less than expected, in part, due to theexistence of downstream activatingmutations, unsuspectedfeedback loops, and signaling pathway cross-talk. As aresult, much effort is currently focused on targeting ofseveral signaling pathways at the same time (1). Cross-talkbetween the epidermal growth factor receptor (EGFR) andthe VEGF signaling pathways plays an important role intumor growth and survival (2). Activation of EGFR signal-ing in tumor cells stimulates the production of VEGF,which then acts in a paracrine fashion on surroundingendothelial cells to stimulate their proliferation and mi-gration (2). Several preclinical studies have combineddifferent EGFR- and VEGF(R)-targeted small-moleculetyrosine kinase inhibitors (TKI) or monoclonal antibodies(mAb) with encouraging results (3–8). Bevacizumab, aVEGF-neutralizing mAb, and cetuximab, an EGFR-targetedmAb, are both approved for treatment of colorectalcancer (CRC). Although an early clinical trial (BOND2)
Authors' Affiliations: 1Laboratory of Cancer Biology and Therapeutics,Centre de Recherche Saint-Antoine; 2Institut National de la Sant�e et de laRecherche M�edicale U938; 3Universit�e Pierre et Marie Curie (University ofParis 6); Departments of 4Pathology and 5Clinical Oncology, Hôpital Saint-Antoine, Paris, France
Note: V. Poindessous, D. Ouaret, and K. El Ouadrani contributed equallyto this work.
Corresponding Author: Annette K. Larsen, Laboratory of Cancer Biologyand Therapeutics, Kourilsky Building 1st floor, Hôpital Saint-Antoine, 184rue du Faubourg Saint Antoine, 75571 Paris, Cedex 12, France. Phone: 33-1-49-28-46-12; Fax: 33-1-42-22-64-29; E-mail: [email protected]
combining bevacizumab and cetuximab looked promising(9), more recent studies (PACCE and CAIRO2) represent-ing almost 1,800 patients showed that the addition ofEGFR-targeted mAbs to bevacizumab plus chemotherapywas no better than bevacizumab plus chemotherapy alone,even in patients with wild-type KRAS tumors (10, 11). Themechanistic basis for these unexpected results is difficult toestablish because no preclinical data are available for thecombination of VEGF- and EGFR-targeted mAbs (12),neither with regard to their activity in xenograft modelsnor with respect to functional biomarkers.A number of recent findings highlight the importance of
intracellular signal transduction in tumor cells. First, it hasbeen shown that the signaling of receptor tyrosine kinases(RTK) such as EGFR and VEGFR is not limited to thereceptors bound to the plasma membrane but that inter-nalized RTKs continue to signal and may even acquirenovel functions (13). Second, several studies have shownthe presence of internal autocrine (intracrine) VEGF/VEGFR1 signaling in different tumor types (14, 15). Thisnotion is further supported by the observation that dele-tion of VEGF-A by homologous recombination, andthus extinction of VEGF/VEGFR intracrine signaling, wasaccompanied by decreased cell growth and increasedspontaneous apoptosis of CRC cells (16). Finally, it hasbeen suggested that autocrine VEGF/VEGFR1 signalingsynergizes with EGFR to promote tumor cell survivaland/or proliferation (17, 18).An important therapeutic implication of these findings is
that strategies to block VEGF or EGFR signaling by inhibi-tion of extracellular ligands or receptors, as is the case forthe mAbs, may only prevent part of the oncogenic signal-ing. In contrast, we would expect that small-molecule TKIsmight be able to interfere with internal RTK signaling andcross-talk, including the VEGF/VEGFR1 intracrine loop. Totest this hypothesis, we selected 2 TKIs including vargatef/BIBF 1120, a triple angiokinase inhibitor of VEGFR,PDGFR, and FGFR (19), and afatinib/BIBW 2992, which
irreversibly inhibits EGFR and HER2 (20). Vargatef iscurrently in phase III trials in non–small cell lung cancer(NSCLC) and ovarian cancer, whereas afatinib has reachedphase III trials for the treatment of NSCLC and breastcancer.
We now report that vargatef and afatinib together showsynergistic activity in CRC models that are refractory tothe bevacizumab and cetuximab combination and eluci-date the mechanistic differences between the TKIs and themAbs. In particular, our results show that only TKIs areable to attenuate intracellular EGFR and VEGFR signaling,which is accompanied by the induction of apoptotic celldeath. Our findings provide a mechanistic explanationfor the failure of the mAbs and indicate that rationallyselected EGFR- and VEGF(R)-targeted agents could becombined for clinical benefit.
Materials and Methods
Xenograft modelsThe antitumor effects of the molecular targeted agents
were evaluated in athymic mice (female NMRI-Foxn1n,6 weeks old) from Taconic bearing HT-29 CRC xenografts.Two million cells were injected into the right flank, and thetreatments were started when the tumors were palpable(median tumor volume �100 mm3). The animals wereweighed daily and the tumor size was determined 3 timesper week. Tumor volumes (mm3) were calculated accord-ing to formula: [(length2 � width)/2]. Boxplot analysis ofthe weights and tumor volumes was carried out using theGraphPad Prism version 5.00 software (GraphPad Soft-ware). Treated/control values were calculated as follows:(average tumor volume of treated animals/average tumorvolume of control animals) � 100. Animals were treatedaccording to institutional guidelines.
ImmunohistochemistryAll biomarker analyses were carried out with tumors
collected after 18-days drug exposure when necrosis isminimal. To measure in vivoDNA synthesis, the thymidineanalogue 5-ethynyl-20-deoxyuridine (EdU; Invitrogen) wasintraperitoneally administered 48 hours before sacrifice(500 mg). The incorporated EdU was revealed by a fluo-rescent-azide coupling reaction (Click-iT; Invitrogen) ofparaffin-embedded tumor samples and counterstainedby 40,6-diamidino-2-phenylindole to reveal the nuclei ofindividual cells. The proportion of apoptotic tumor cellswas scored by the terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) assay (In SituCell Death Detection kit; Roche Applied Science).
The following antibodies were used for immunohisto-chemistry analysis: anti-phospho-EGFR antibodies (#12351;Santa Cruz), which recognize Tyr1173-phosphorylatedEGFR, anti-phospho-VEGFR1 antibodies (#07-758; Milli-pore), which recognize Tyr1213-phosphorylated VEGFR1,and the relevant Cy3-conjugated secondary antibodies(Jackson ImmunoResearch). All images were captured witha fluorescence microscope, and the fluorescence intensities
Translational Relevance
Both epidermal growth factor receptor (EGFR)- andVEGF-directed monoclonal antibodies are approved fortreatment of colorectal cancer. However, their combi-nation has shown disappointing clinical activity despitethe extensive cross-talk between EGFR and VEGF(R)signaling pathways. This article reports that combina-tions of EGFR- and VEGF(R)-targeted small-moleculetyrosine kinase inhibitors (TKI) inhibit intracrine sig-naling, trigger apoptotic cell death, and show synergisticantitumor activity in colorectal cancer xenografts andcells, independent of KRAS status. In contrast, combi-nations of monoclonal antibodies targeting the samepathways were exclusively cytostatic, with no moreactivity than either drug alone. This work is highlytranslational and provides a rationale for clinical trialsof the TKIs, even in patients with mutant KRAS.
Synergistic Activity of TKIs Targeting EGFR and VEGF(R)
www.aacrjournals.org Clin Cancer Res; 17(20) October 15, 2011 6523
were determined by the MetaMorph software (UniversalImaging Corporation) for quantitative analysis.
For the quantitative analysis of the in vivoDNA synthesis(EdU incorporation), the data represent the ratio betweenEdU-positive cells and the total number of viable cells andare the averages of 5 fields per tumor (each field represent-ing approximately 1,700 cells) from 3 different tumors. Forthe quantitative determination of apoptosis, the data rep-resent the ratio between TUNEL-positive apoptotic cellsand the total area of viable cells and are the averages of 5fields per tumor for 4 different tumors. For the quantitativeanalysis of the signal intensity for phospho-EGFR andphospho-VEGFR1, the data represent the average fluores-cence intensity of treated tumors, compared with the treat-ment intensity of control tumors, and are the averages of 5fields per tumor for 4 different tumors.
Tumor cells, cytotoxicity assays, and flowcytometric analysis
Tumor cells were kindly provided by Richard Camalier(Division of Cancer Treatment and Diagnosis TumorRepository, National Cancer Institute) and by RichardHamelin (Paris, France). Cellular viability was determinedby the MTT viability assay after 120-hour continuous drugexposure as described previously (21). Cell-cycle analysiswas carried out as described (22), whereas the proportionof apoptotic cells was characterized by flow cytometryusing the APO-BRDU Kit from BD Biosciences. Drug com-bination effects were determined by the analysis of Chouand Talalay based on the median-effect equation (23, 24)and are indicated in terms of combination index. Data wereanalyzed by using the concentration effect analysis software(Biosoft). Statistical analysis and graphs were accom-plished by GraphPad Prism version 5.00 (GraphPadSoftware).
Immunoblot analysisImmunoblot analysis was carried out as described pre-
viously (25). The primary antibody was directed againstp27Kip1 (#3686; Cell Signaling), whereas the secondaryantibody was purchased from Jackson ImmunoResearch.
Results
Influence of vargatef, afatinib, and their combinationon tumor growth in comparison with bevacizumaband cetuximab
For these experiments, we used HT-29 xenografts, whichis a classical CRCmodel for the testing of anticancer agents.The animals were exposed to doses that lead to comparableantitumor activity of the 4 agents when given alone. Nodrug-associated morbidity or mortality was observed forany of the groups.
Simultaneous exposure to bevacizumab or cetuximabwas no better than either agent alone (Fig. 1A and C). Incontrast, the combination of vargatef and afatinib wasassociated with a clear synergistic effect compared witheither agent alone (Fig. 1B and C).
Influence of vargatef, afatinib, and their combinationon tumor proliferation and viability compared withbevacizumab and cetuximab combinations
Tumor growth inhibition can be due to cytostatic (cell-cycle related) or cytotoxic effects. To distinguish betweenthese possibilities, we first determined the influence of the4 agents on in vivo DNA synthesis. Tumor-bearing mice
Days
A
BDays
Tu
mo
r vo
lum
e (
mm
3)
Tu
mo
r vo
lum
e (
mm
3)
Tu
mo
r vo
lum
e (
mm
3)
0
200
400
600
800
1,000
1,200
2826242119171410750
0
200
400
600
800
1,000
1,200
2826242119171410750
C
0
500
1,000
1,500
Figure 1. A, nude mice with HT-29 CRC xenografts were dosed withvehicle (&), bevacizumab (Beva) at 5 mg/kg i.p. every 3 days ( ),cetuximab (Cetux) at 1 mg per mouse i.p. every 3 days ( ), or bevacizumaband cetuximab together ( ). The curves represent the average tumorgrowth of at least 7 animals per group. B, tumor-bearing nude mice weredosed with vehicle (&), vargatef (Varga) at 10 mg/kg per os once daily ( ),afatinib (Afa) at 12.5 mg/kg per os once daily ( ), or vargatef andafatinib together ( ). The curves represent the average tumor growth of atleast 7 animals per group. C, box and whisker plot of tumor volumes inmice with HT-29 xenografts after 4-week treatment with bevacizumab,cetuximab, vargatef, afatinib, or their combinations. Lines, medians;boxes, 25th to 75th percentile interquartile ranges; and whiskers, thehighest and lowest values for a given treatment. Dashed horizontal line,the median value of the vehicle group.
Poindessous et al.
Clin Cancer Res; 17(20) October 15, 2011 Clinical Cancer Research6524
were injected with EdU, a thymidine analogue, 48 hoursbefore sacrifice, and the incorporation was determined by afluorescent-azide coupling reaction. Quantitative imageanalysis showed that treatment with bevacizumab, cetux-imab, or vargatef alone was accompanied by approximately40% inhibition of the in vivo DNA synthesis, whereasafatinib inhibited the DNA synthesis by almost 70%(Fig. 2A). Bevacizumab and cetuximab together or vargatefand afatinib together was no better than the most active ofthe 2 agents when given alone (Fig. 2A).
Next, the TUNEL assay was used to determine the influ-ence of the 4 agents on apoptotic tumor cell death. Treat-ment with bevacizumab, cetuximab, or vargatef alone wasnot accompanied by any notable increase in apoptosiscompared with control tumors, whereas exposure to afa-tinib doubled the fraction of apoptotic cells (Fig. 2B).Simultaneous exposure to bevacizumab and cetuximabwas not accompanied by increased apoptosis, whereassimultaneous exposure to vargatef and afatinib increasedthe fraction of apoptotic cells almost 5-fold, comparedwith vehicle-treated control tumors (Fig. 2B).
Influence of prolonged drug exposure on thephosphorylation of EGFR and VEGFR1
Survival of CRC cells has been linked to internalautocrine (intracrine) signaling. We therefore deter-mined the influence of prolonged drug exposure onreceptor autophosphorylation by quantitative immuno-histochemistry. HT-29 control tumors displayed a strongsignal for phospho-EGFR, which was both membrane-associated and intracellular (Fig. 3A). Exposure tobevacizumab plus cetuximab had modest influence onthe intensity (86% � 7%, compared with vehicle) ofphospho-EGFR and no effect on its distribution. In con-trast, exposure to vargatef plus afatinib was accompani-ed by a pronounced diminution of the total phospho-EGFR signal (36% � 4%, compared with vehicle) anda reduction of the intracellular fraction (Fig. 3A).
The presence of tumor-associated VEGFR1 has beenreported for CRC cells (15, 26) and for tumors in patientswith CRC (27). In agreement, control tumors displayed astrong signal for phospho-VEGFR1, which was both mem-brane-associated and intracellular (Fig. 3B). Exposure tobevacizumab plus cetuximab was accompanied by a decre-ase of the phospho-VEGFR1 signal to 81% � 7% of vehiclecontrols without any detectable alterations of the cellulardistribution. Similar exposure to vargatef plus afatinib wasaccompanied by a decrease to 52% � 8% of controls anda marked reduction of the intracellular fraction (Fig. 3B).
Influence of afatinib and vargatef on the viability ofcolorectal carcinoma cells
For further characterization, the activity of afatinib to-ward a CRC cell panel was determined. For comparison, weincluded 3 reference cell lines expressing high levels ofEGFR and/or HER2 including EGFR-overexpressing humanepidermoid A431 carcinoma cells, HER2-overexpressingNCI-N87 gastric carcinoma cells, and HER2-overexpressingBT-474 breast carcinoma cells. The results revealed a 130-fold range in the sensitivity to afatinib (Fig. 4A), with IC50
values (drug concentration inhibiting cell growth by 50%compared with untreated controls) ranging from 0.05 to6.5 mmol/L and an average IC50 of 1.9 mmol/L.
Tumor-associated VEGFR1 is believed to influence cel-lular survival and/or proliferation (14, 15, 18). In agree-ment, our results show that vargatef reduced the viability ofCRC carcinoma cells, with IC50 values ranging from 0.6 to4.5 mmol/L and an average IC50 of 2.2 mmol/L (Fig. 4B).
AVe B + C V + A
V + AB + CVeB
0
5
10
15
20
25
Pro
life
rati
ng
cells
(% o
f co
ntr
ol)
TU
NE
L
(% t
ota
l are
a)
0
20
40
60
80
100
Figure 2. A, influence of bevacizumab, cetuximab, vargatef, afatinib, andtheir combinations on the in vivo DNA synthesis of HT-29 tumors asmeasured by EdU incorporation. Top, representative images of tumorsfrom mice treated with vehicle (Ve), bevacizumab (Beva) and cetuximab(Cetux) together (BþC), or vargatef (Varga) and afatinib (Afa) together (VþA). Tumor cell nuclei are indicated in blue, whereas nuclei with active DNAsynthesis appear in white. Bottom, the data represent the ratio betweenEdU-positive cells and the total number of viable cells and are the averagesof 5 fields per tumor (each field representing �1,700 cells) from 3 differenttumors. All values are normalized with respect to the control tumors(vehicle), which were considered as 100%. B, influence of bevacizumab,cetuximab, vargatef, afatinib, and their combinations on the induction ofapoptotic cell death in HT-29 tumors as measured by the TUNEL assay.Top, representative images of tumors from animals treated with vehicle(Ve), bevacizumab plus cetuximab together (B þ V), or vargatef plusafatinib together (V þ A). Tumor cell nuclei are indicated in blue, whereasapoptotic cells appear in white. Bottom, the data represent the ratiobetween TUNEL-positive apoptotic cells and the total area of viable cellsand are the averages of 5 fields per tumor for 4 different tumors.
Synergistic Activity of TKIs Targeting EGFR and VEGF(R)
www.aacrjournals.org Clin Cancer Res; 17(20) October 15, 2011 6525
To determine whether the observed effects were drug-specific or rather reflected the intrinsic sensitivity of theindividual cell lines, the IC50 values for vargatef wereplotted against the IC50 values for afatinib (Fig. 4C). Dataanalysis by the Student t test revealed no correlation (r2 ¼0.14, P ¼ 0.22) between vargatef and afatinib, confirmingthat the sensitivity to the 2 drugs is mediated by differentpathways.
Influence of vargatef, afatinib, and their combinationon cell-cycle progression
The growth inhibitory effects of vargatef and afatinibin vitro could be due to cytostatic or cytotoxic effects. Cell-cycle analysis of HT-29 and LS513 cells showed that bothvargatef and afatinib induced a pronounced cell-cyclearrest in G1 by 24 hours that lasted throughout the120-hour incubation period. Interestingly, simultaneousexposure to both drugs was only associated with a mar-ginally increased G1 fraction compared with either agentalone (Fig. 5A and B).
Cell-cycle arrest of erlotinib-treated lung cancer cellshas been causally linked to the induction of the cyclin-dependent kinase inhibitor p27Kip1 (28, 29). Our resultsshow that p27Kip1 is also induced in CRC cells exposedto vargatef and/or afatinib (Fig. 5C).
Influence of vargatef, afatinib, and their combinationon the viability of colorectal carcinoma cells withdifferent KRAS and BRAF status
Induction of apoptotic cell death was determined by theTUNEL assay. Continued exposure of HT-29 cells to var-gatef or afatinib as a single agent was accompanied byinduction of apoptotic cell death in at least 10% of the cellsafter 96-hour drug exposure (Fig. 6A). In contrast, noincreased cell death was observed for LS513 cells through-out the 120-hour incubation period (Fig. 6B). Simulta-neous exposure to vargatef and afatinib was accompaniedby a marked increase of apoptotic HT-29 cells after72 hours, which reached more than 40% of the total by120 hours (Fig. 6A). Unexpectedly, simultaneous exposure
Figure 3. Influence of prolongedexposure to vehicle, bevacizumab(Beva) plus cetuximab (Cetux)together, or vargatef (Varga) plusafatinib (Afa) together on (A)phosphorylated EGFR (P-EGFR)and (B) phosphorylated VEGFR1(P-VEGFR1) in HT-29 tumors. Thedistribution of phospho-EGFR orphospho-VEGFR1 is shown in red,whereas the nuclei appear in blue.Bottom, quantitative analysis ofthe intensity of the signal forphospho-EGFR and phospho-VEGFR1. The data represent theaverage fluorescence intensity oftreated tumors compared with thetreatment intensity of controltumors and are the averages of 5fields per tumor for 4 differenttumors. Bars, SD.
A BP-EGFR
Varg
a +
Afa
Veh
icle
Beva +
Cetu
x
0
50
100
P-V
EG
FR
1 i
nte
nsit
y
(% o
f co
ntr
ol)
0
50
100
P-E
GF
R i
nte
nsit
y
(% o
f co
ntr
ol)
P-VEGFR1
Varg
a +
Afa
Veh
icle
Beva +
Cetu
x
Poindessous et al.
Clin Cancer Res; 17(20) October 15, 2011 Clinical Cancer Research6526
to vargatef and afatinib also induced apoptosis in at least20% of LS513 cells (Fig. 6B). In confirmation, the analysisof Chou and Talalay of LS513 cells exposed to differentconcentrations of vargatef and afatinib showed at leastadditive activity, except at low doses. At drug combinationsleading to more than 50% loss of viability, the combina-tion of the 2 drugs was synergistic (Fig. 6C).To extend these findings, the influence of vargatef and
afatinib was determined for a CRC cell panel with differentKRAS or BRAF status. The results show that the vargatef andafatinib combination was more cytotoxic than either drugalone for 8 of 8 cell lines tested, regardless of KRAS andBRAF mutational status, or whether the cells displayed themicrosatellite instability phenotype (MSI/MIN) or loss ofheterozygosity (LOH/CIN/MSS; Fig. 6D).
Discussion
This study was undertaken to determine whether thedisappointing results in recent clinical trials with combina-tions of EGFR- and VEGF-targetedmAbs could be explainedby their limited activity on intracellular signaling events.Although several preclinical studies have previously com-bined different VEGF(R)- and EGFR-targeted agents, thepresent study is, to the best of our knowledge, the only tohave compared the activity of TKIs with mAbs in the samein vivo model.
We found that vargatef and afatinib together showedstrong tumor growth inhibitory activity toward HT-29 CRCxenografts, compared with either drug alone, which wasassociated with increased tumor cell death. In comparison,bevacizumab and cetuximab together were no more activethan either drug alone and showed exclusively cytostaticactivity.
Little is known about how prolonged drug exposureinfluence RTK autophosphorylation and thus their activity.Only TKIs are likely to inhibit intracellular RTKs following
A
B
IC5
0 v
arg
ate
f (µ
mo
l/L
)
IC50 vargatef (µmol/L)
0
1
2
3
4
5
C
01
2
3
4
5r 2 = 0.141
P = 0.22
76
543210
6
IC5
0 a
fati
nib
(µ
mo
l/L
)
IC50 a
fati
nib
(µ
mo
l/L
)
012345
7
SW48
Lovo
SW1116
SW 620
SW620
SW 48
Figure 4. A, the growth inhibitory activity of afatinib (white columns) or(B) vargatef (dark columns) toward CRC cells was determined by the MTTviability assay after 120 hours of continuous drug exposure and isexpressed as IC50 values (drug concentration inhibiting cell growth by50% compared with untreated controls). The hatched columnscorrespond to reference cell lines with overexpression of EGFR (A431) orHER2 (NCI-N87, BT-474). All values are averages of at least 3 independentexperiments each done in duplicate. C, the IC50 values for vargatef wereplotted against the IC50 values for afatinib for each cell line. The resultsrevealed no correlation between the sensitivity to the 2 agents.
A
B
Cells i
n G
1
(% o
f to
tal)
Ce
lls
in
G1
(% o
f to
tal)
50
60
70
80
90
120967248240
Incubation time (h)
50
60
70
80
90
120967248240Incubation time (h)
C Vargatef Afatinib
Vargatef
+ Afatinib
p27KIP1
7248244NT 7248244 7248244 (h)
β-Actin
Figure 5. The influence of vargatef ( ), afatinib ( ), or their combination( ) on the proportion of (A) HT-29 or (B) LS513 cells in the G1 phase of thecell cycle. Cells were exposed to the IC50 dose of each drug, and the cell-cycle distribution was determined by flow cytometric analysis. All valuesare averages of 2 independent experiments each done in duplicate. Bars,SD. C, HT-29 cells were exposed to the IC50 dose of each drug, and theprotein levels of the cell-cycle regulator p27Kip1 were determined byimmunoblot analysis. The protein levels of b-actin are shown as theloading control. NT, not treated.
Synergistic Activity of TKIs Targeting EGFR and VEGF(R)
www.aacrjournals.org Clin Cancer Res; 17(20) October 15, 2011 6527
short-term exposure. However, because receptor internal-ization and stability are also influenced by the phosphor-ylation status, it was possible that long-term exposure toboth TKIs and mAbs could modify the levels and cellulardistribution of active, phosphorylated RTKs. Controltumors displayed both membrane-associated and intracel-lular phospho-EGFR and phospho-VEGFR1. Prolongedexposure to bevacizumab plus cetuximab hadmodest effecton the levels of phospho-EGFR in HT-29 tumors and nodetectable influence on the distribution. Similar exposureto vargatef plus afatinib was accompanied by almost 65%reduction of the phospho-EGFR signal and a reduction ofthe intracellular fraction. Bevacizumab plus cetuximabexposure was accompanied by a 20% diminution of thephospho-VEGFR1 signal and no detectable influence on itsdistribution, whereas vargatef plus afatinib exposureresulted in a 50% diminution of the phospho-VEGFR1signal and a marked reduction of the intracellular fraction.These results indicate that prolonged exposure to vargatefplus afatinib together reduces the intracellular levels of
both phospho-VEGFR1 and phospho-EGFR, whereas sim-ilar exposure to bevacizumab and cetuximab combinationshas no detectable influence. The close association betweenattenuation of intracellular phospho-EGFR and phospho-VEGFR1 and the induction of apoptotic cell death in thesein vivo models is in line with recent results for cellularmodels that document the important contribution of in-tracellular signaling for tumor cell survival (13–18).
Further characterization of vargatef and afatinib in a CRCcell panel revealed that prolonged exposure to both com-pounds was accompanied by decreased cellular viability.The cytotoxic activity of vargatef toward CRC cells is inapparent contrast to our in vivo findings, where vargatefalone showed only cytostatic activity. However, it shouldbe noted that the vargatef concentration used for the in vivostudies was adapted to be isoeffective with the 3 otheragents and is far below the maximal tolerated dose.
Vargatef or afatinib alone induced a prolonged G1 arrestin both LS513 and HT-29 cells, which was associated withupregulation of the cyclin-dependent kinase inhibitor
C
A
724824 120960
10
20
30
40
50
TU
NE
L-p
os
itiv
e c
ell
s (
%)
Exposure time (h)
0
10
20
30
40
50
12096724824TU
NE
L-p
os
itiv
e c
ell
s (
%)
Exposure time (h)
B
Co
mb
inati
on
in
dex
Fraction affected
0.000
1
2
0.25 0.50 0.75 1.00 0
25
50
75
Via
bil
ity (
%)
LS
513
SW
48
FE
T
HC
T-1
16
HT
-29
Lo
Vo
SW
480
DL
D-1
KRAS mut wt
BRAF
wtmut mut
wt
mut
wt
mut
wt
mut
wtmut wtwtwt
LOH/MSI MSIMSIMSIMSILOHLOHLOHLOH
D
Figure 6. The influence of vargatef (dark columns), afatinib (white columns), or their combination (gray columns) on the proportion of apoptotic cells for HT-29(A) or LS513 (B) cells. Cells were exposed to the IC50 dose of each drug, and the proportion of TUNEL-positive apoptotic cells was determined by flowcytometric analysis. All values are averages of 2 independent experiments each done in duplicate. Bars, SD. C, LS513 cells were exposed to increasingconcentrations of vargatef in combination with afatinib at concentrations corresponding to the IC10 ( ), IC20 ( ), IC30 ( ), IC40 ( ), and IC50 ( ) doses.The cytotoxic effects were analyzed according to the Chou and Talalay method. Combination index values of less than 0.8 indicate synergy, values between0.8 and 1.2 indicate additive effects, and values of more than 1.2 indicate antagonism. D, the influence of vargatef (dark columns), afatinib (white columns),and their combination (gray columns) on the viability of 8 different CRC cell lines was determined by the MTT assay. All values are averages of at least2 independent experiments each done in duplicate. Bars, SD. mut, mutant; wt, wild-type.
Poindessous et al.
Clin Cancer Res; 17(20) October 15, 2011 Clinical Cancer Research6528
p27Kip1. Vargatef and afatinib together did not result in anyfurther enrichment of G1 phase cells compared with eitheragent alone, coherent with the in vivo findings. Therefore,one possible explanation for why combinations of EGFRand VEGF(R)-targeted molecules are no better in inhibitingproliferation than the most active of the two when givenalone may be that both compounds depend on the activityof the same cell-cycle mediator.In contrast, vargatef and afatinib together increased the
fraction of apoptotic cells, which was particularly strikingfor LS513 cells, where either drug alone was incapable ofinducing apoptosis. Accordingly, Chou and Talalay anal-ysis of LS513 cells exposed to different combinations ofvargatef and afatinib showed mostly additive to synergisticeffects.Combinations of vargatef and afatinib were associated
with at least additive effects in 8 of 8 CRC models tested,independent of KRAS and BRAF mutational status or themicrosatellite instability (MSI/MIN) phenotype. In agree-ment, it has been reported that the vargatef plus afatinibcombination showed activity in mice with Ras-dependentsarcomas (18). Mutation of KRAS or BRAF is a negativepredictive factor for EGFR-targeted antibodies in patientswith CRC (30), whereas the influence of KRAS mutationsis less clear for the TKIs (31). Interestingly, mutant KRASis also a negative predictive factor for inhibitors of themTOR and this is observed in both cellular and xenograftmodels as well as in patients with cancer (32), suggestingthat cellular models may be useful for establishing theinfluence of KRAS status on the response to targetedagents. Therefore, the activity of the vargatef and afatinibcombination toward CRC models with mutant KRAS orBRAF is an important observation that merits clinicalvalidation considering that up to 40% of patients withCRC have mutant KRASwhereas approximately 10% havemutant BRAF.Because both vargatef and afatinib are multi-targeted
agents, one could ask to which extend the activity of thesecompounds depends on the inhibition of VEGFR1 andEGFR. This question has recently been addressed for K5-SOS mice with epidermal carcinomas where keratinocyte-specific deletion of the genes for VEGF and EGFR hadcomparable influence on tumor growth as pharmacologicinhibition of VEGFR and EGFR signaling by vargatef andafatinib (18). Thus, it seems that inhibition of VEGF(R)and EGFR signaling is an essential contributor to theactivity of the vargatef plus afatinib combination. However,it is likely that the additional targets of the 2 drugs alsocontribute to the antitumor activity. Indeed, systembiologymodels predict that evolvable systems such as RTK net-works are resistant to interception of individual compo-nents but fragile when subjected to multiple simultaneousperturbations (33), as would be the case for vargatef andafatinib.Another major question is to which extend the findings
presented here are applicable for combinations of otherVEGF(R)- and EGFR-directed agents. The biological activityof all TKIs depends on multiple factors including the
specificity, the degree, and the duration of target inhibition.Both vargatef and afatinib are multi-targeted agents that areassociated with strong RTK inhibition for prolonged per-iods of time (19, 20). Therefore, although it might befeasible to combine other VEGF(R) and EGFR-targetedagents besides the 2 compounds described here, it isunlikely that all combinations of 2 TKIs, or of 1 TKI with1 mAb, will be active.
The toxic side effects of combining VEGF(R)- and EGFR-targeted agents are also likely to depend, at least in part, onthe properties of the individual compounds. A recent phaseI study concluded that continuous daily oral exposure toafatinib was safe and had durable antitumor activity (34)whereas 2 phase I studies reported that continuous vargatefdisplays a favorable safety and pharmacokinetics profilewith first efficacy signals (35, 36). Furthermore, it has beenshown that the 2 drugs could be combined in patients withCRC, even if the scheduling and duration remains to beestablished (37).
Monoclonal antibodies directed against growth factorreceptors and their ligands principally target the humanform of these proteins, for which reason some of theactivity of the mAbs may be underestimated in xenograftmodels. However, we do not believe this is an issue for thepresent work, because the lack of synergistic activity for the2mAb is in line with the clinical findings. Furthermore, thislimitation does not apply to our principal finding thatelucidates the differential effects of TKI and mAb combina-tions on intracellular EGFR and VEGF(R) signaling in thetumor cells, which are of human origin.
Taken together, our findings underline the utility ofpreclinical models and the use of functional biomarkersfor the testing of molecular-targeted anticancer agents,alone or in combination. The results elucidate a mecha-nistic reason for the clinical failure of the 2 mAbs andprovide a rationale for trials of the TKIs, even in patientswith mutant KRAS.
Disclosure of Potential Conflicts of Interest
The sponsors had no role in the study design, data collection andanalysis, interpretation of the results, the preparation of the manuscript,or the decision to submit the manuscript for publication.
Acknowledgments
The authors thank the contribution of Isabelle Renault, Delphine Muller,and Tatiana Ledent from the animal facilities at Saint-Antoine ResearchCenter and Anne-Marie Faussat at the flow cytometric platform of IFR65.
Grant Support
This work was financed in part by Boehringer Ingelheim (A.K. Larsen) and byAssociation pour la Recherche sur le Cancer (ARC), Villejuif, France. P. Boudouwas supported by a fellowship from Foundation Soldati.
The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicate thisfact.
Received June 21, 2011; revised August 15, 2011; accepted August 16,2011; published OnlineFirst August 31, 2011.
Synergistic Activity of TKIs Targeting EGFR and VEGF(R)
www.aacrjournals.org Clin Cancer Res; 17(20) October 15, 2011 6529
References1. Gossage L, Eisen T. Targeting multiple kinase pathways: a change in
paradigm. Clin Cancer Res 2010;16:1973–8.2. Larsen AK, Ouaret D, El Ouadrani K, Petitprez A. Targeting EGFR and
VEGF(R) pathway cross-talk in tumor survival and angiogenesis.Pharmacol Ther 2011;131:80–90.
3. Ciardiello F, Bianco R, Damiano V, Fontanini G, Caputo R, PomaticoG, et al. Antiangiogenic and antitumor activity of anti-epidermalgrowth factor receptor C225 monoclonal antibody in combinationwith vascular endothelial growth factor antisense oligonucleotide inhuman GEO colon cancer cells. Clin Cancer Res 2000;6:3739–47.
4. Shaheen RM, Ahmad SA, Liu W, Reinmuth N, Jung YD, Tseng WW,et al. Inhibited growth of colon cancer carcinomatosis by antibodies tovascular endothelial and epidermal growth factor receptors. Br JCancer 2001;85:584–9.
5. Sini P,Wyder L, Schnell C, O’Reilly T, Littlewood A, Brandt R, et al. Theantitumor and antiangiogenic activity of vascular endothelial growthfactor receptor inhibition is potentiated by ErbB1 blockade. ClinCancer Res 2005;11:4521–32.
6. Yokoi K, Thaker PH, Yazici S, Rebhun RR, Nam DH, He J, et al. Dualinhibition of epidermal growth factor receptor and vascular endothelialgrowth factor receptor phosphorylation by AEE788 reduces growthand metastasis of human colon carcinoma in an orthotopic nudemouse model. Cancer Res 2005;65:3716–25.
7. Tonra JR, Deevi DS, Corcoran E, Li H, Wang S, Carrick FE, et al.Synergistic antitumor effects of combined epidermal growth factorreceptor and vascular endothelial growth factor receptor-2 targetedtherapy. Clin Cancer Res 2006;12:2197–207.
9. Saltz LB, Lenz HJ, Kindler HL, Hochster HS, Wadler S, Hoff PM, et al.Randomized phase II trial of cetuximab, bevacizumab, and irinotecancompared with cetuximab and bevacizumab alone in irinotecan-re-fractory colorectal cancer: the BOND-2 study. J Clin Oncol2007;25:4557–61.
10. Hecht JR, Mitchell E, Chidiac T, Scroggin C, Hagenstad C, Spigel D,et al. A randomized phase IIIB trial of chemotherapy, bevacizumab,and panitumumab compared with chemotherapy and bevacizumabalone for metastatic colorectal cancer. J Clin Oncol 2009;27:672–80.
11. Tol J, Koopman M, Cats A, Rodenburg CJ, Creemers GJ, SchramaJG, et al. Chemotherapy, bevacizumab, and cetuximab in metastaticcolorectal cancer. N Engl J Med 2009;360:563–72.
12. Punt CJ, Tol J. More is less - combining targeted therapies inmetastatic colorectal cancer. Nat Rev Clin Oncol 2009;6:731–3.
13. Lemmon MA, Schlessinger J. Cell signaling by receptor tyrosinekinases. Cell 2010:141;1117–34.
14. Lee TH, Seng S, SekineM, Hinton C, Fu Y, AvrahamHK, et al. Vascularendothelial growth factor mediates intracrine survival in human breastcarcinoma cells through internally expressed VEGFR1/FLT1. PLoSMed 2007;4,e186.
15. Naik S, Dothager RS, Marasa J, Lewis CL, Piwnica-Worms D. Vas-cular endothelial growth factor receptor-1 is synthetic lethal to aber-rant {beta}-catenin activation in colon cancer. Clin Cancer Res2009;15:7529–37.
16. Samuel S, Fan F, Dang LH, Xia L, Gaur P, Ellis LM. Intracrine vascularendothelial growth factor signaling in survival and chemoresistance ofhuman colorectal cancer cells. Oncogene 2011;30:1205–12.
17. Bianco R, Rosa R, Damiano V, Daniele G, Gelardi T, Garofalo S, et al.Vascular endothelial growth factor receptor-1 contributes to resis-tance to anti-epidermal growth factor receptor drugs in human cancercells. Clin Cancer Res 2008;14:5069–80.
18. Lichtenberger BM, Tan PK, Niederleithner H, Ferrara N, Petzelbauer P,Sibilia M. Autocrine VEGF signaling synergizes with EGFR in tumorcells to promote epithelial cancer development. Cell 2010;140:268–79.
19. Hilberg F, Roth GJ, Krssak M, Kautschitsch S, Sommergruber W,Tontsch-Grunt U, et al. BIBF 1120: triple angiokinase inhibitor with
sustained receptor blockade and good antitumor efficacy. Cancer Res2008;68:4774–82.
20. Li D, Ambrogio L, Shimamura T, Kubo S, Takahashi M, Chirieac LR,et al. BIBW2992, an irreversible EGFR/HER2 inhibitor highly effectivein preclinical lung cancer models. Oncogene 2008;27:4702–11.
21. Soares DG, Escargueil AE, Poindessous V, Sarasin A, de GramontA, Bonatto D, et al. Replication and homologous recombinationrepair regulate DNA double-strand break formation by the antitumoralkylator ecteinascidin 743. Proc Natl Acad Sci U S A 2007;104:13062–7.
22. Skladanowski A, Côme MG, Sabisz M, Escargueil AE, Larsen AK.Down-regulation of DNA topoisomerase IIalpha leads to prolongedcell cycle transit in G2 and early M phases and increased survival tomicrotubule-interacting agents. Mol Pharmacol 2005;68:625–34.
23. Chou TC. Drug combination studies and their synergy quantificationusing the Chou-Talalay method. Cancer Res 2010;70:440–6.
24. Poindessous V, Koeppel F, Raymond E, Cvitkovic E, Waters SJ,Larsen AK. Enhanced antitumor activity of irofulven in combinationwith 5-fluorouracil and cisplatin in human colon and ovarian carcino-ma cells. Int J Oncol 2003;23:1347–55.
25. Escargueil AE, Poindessous V, Soares DG, Sarasin A, Cook PR,Larsen AK. Influence of irofulven, a transcription-coupled repair-spe-cific antitumor agent, on RNA polymerase activity, stability anddynamics in living mammalian cells. J Cell Sci 2008;121:1275–83.
26. Fan F, Wey JS, McCarty MF, Belcheva A, Liu W, Bauer TW, et al.Expression and function of vascular endothelial growth factor re-ceptor-1 on human colorectal cancer cells. Oncogene 2005;24:2647–53.
27. Bates RC, Goldsmith JD, Bachelder RE, BrownC, ShibuyaM, OettgenP, et al. Flt-1-dependent survival characterizes the epithelial-mesen-chymal transition of colonic organoids. Curr Biol 2003;13:1721–7.
28. Busse D, Doughty RS, Ramsey TT, Russell WE, Price JO, FlanaganWM, et al. Reversible G(1) arrest induced by inhibition of the epidermalgrowth factor receptor tyrosine kinase requires up-regulation of p27(KIP1) independent of MAPK activity. J Biol Chem 2000;275:6987–95.
29. Ling YH, Li T, Yuan Z, Haigentz Jr M, Weber TK, Perez-Soler R.Erlotinib, an effective epidermal growth factor receptor tyrosine kinaseinhibitor, induces p27KIP1 up-regulation and nuclear translocation inassociation with cell growth inhibition and G1/S phase arrest in humannon-small-cell lung cancer cell lines. Mol Pharmacol 2007;72:248–58.
30. Siena S, Sartore-Bianchi A, Di Nicolantonio F, Balfour J, Bardelli A.Biomarkers predicting clinical outcome of epidermal growth factorreceptor-targeted therapy in metastatic colorectal cancer. J NatlCancer Inst 2009;101:1308–24.
31. Roberts PJ, Stinchcombe TE, Der CJ, Socinski MA. Personalizedmedicine in non-small-cell lung cancer: is KRAS a useful marker inselecting patients for epidermal growth factor receptor-targetedtherapy? J Clin Oncol 2010;28:4769–77.
32. Di Nicolantonio F, Arena S, Tabernero J, Grosso S, Molinari F,Macarulla T, et al. Deregulation of the PI3K and KRAS signalingpathways in human cancer cells determines their response to ever-olimus. J Clin Invest 2010;120:2858–66.
33. Amit I, Wides R, Yarden Y. Evolvable signaling networks of receptortyrosine kinases: relevance of robustness tomalignancy and to cancertherapy. Mol Syst Biol 2007;3:151.
34. Yap TA, Vidal L, Adam J, Stephens P, Spicer J, Shaw H, et al. Phase Itrial of the irreversible EGFR and HER2 kinase inhibitor BIBW 2992 inpatients with advanced solid tumors. J Clin Oncol 2010;28:3965–72.
35. Mross K, Stefanic M, Gmehling D, Frost A, Baas F, Unger C, et al.Phase I study of the angiogenesis inhibitor BIBF 1120 in patients withadvanced solid tumors. Clin Cancer Res 2010;16:311–9.
36. Okamoto I, Kaneda H, Satoh T, Okamoto W, Miyazaki M, Morinaga R,et al. Phase I safety, pharmacokinetic, and biomarker study of BIBF1120, an oral triple tyrosine kinase inhibitor in patients with advancedsolid tumors. Mol Cancer Ther 2010;9:2825–33.
37. Bouch�e O, DucreuxM, Lledo G, Andr�e T, Maildrault-Goebel F, StopferP, et al. A phase II trial of weekly alternating sequential administrationof BIBF1120 and BIBW2992 in patients with advanced colorectalcancer. J Clin Oncol 26, 2008(suppl; abstr 15001).
Poindessous et al.
Clin Cancer Res; 17(20) October 15, 2011 Clinical Cancer Research6530
2011;17:6522-6530. Published OnlineFirst August 31, 2011.Clin Cancer Res Virginie Poindessous, Djamila Ouaret, Karima El Ouadrani, et al. Monoclonal Antibodies
ofActivity in Colorectal Cancer Models Refractory to Combinations EGFR- and VEGF(R)-Targeted Small Molecules Show Synergistic
Updated version
10.1158/1078-0432.CCR-11-1607doi:
Access the most recent version of this article at:
![VascularEndothelialGrowthFactorPlusEpidermal ... · colon cancer cells [29]. In a preclinical model of gastric cancer, inhibition of VEGF and EGFR signaling resulted in significantly](https://static.documents.pub/doc/80x56/60c5459c68257f28be42ee1c/vascularendothelialgrowthfactorplusepidermal-colon-cancer-cells-29-in-a-preclinical.jpg)
