Preclinical Activity of Abemaciclib Alone or in ... · Neil O'Brien1, Dylan Conklin1, Richard Beckmann2,Tong Luo1, Kevin Chau1, Josh Thomas 1 , Ann Mc Nulty 2 , Christophe Marchal

Small Molecule Therapeutics

Preclinical Activity of Abemaciclib Alone or inCombination with Antimitotic and TargetedTherapies in Breast CancerNeil O'Brien1, Dylan Conklin1, Richard Beckmann2, Tong Luo1, Kevin Chau1,Josh Thomas1, Ann Mc Nulty2, Christophe Marchal2, Ondrej Kalous1,Erika von Euw1, Sara Hurvitz1, Colleen Mockbee2, and Dennis J. Slamon1

Abstract

The cyclinD:CDK4/6:Rb axis is dysregulated in a variety ofhuman cancers. Targeting this pathway has proven to be a suc-cessful therapeutic approach in ERþ breast cancer. In this study,in vitro and in vivo preclinical breast cancer models were used toinvestigate the expanded use of the CDK4/6 inhibitor, abemaci-clib. Using a panel of 44 breast cancer cell lines, differentialsensitivity to abemaciclib was observed and was seen predomi-nately in the luminal ERþ/HER2� and ERþ/HER2þ subtypes.However, a subset of triple-negative breast cancer (TNBC) celllines with intact Rb signaling were also found to be responsive.Equivalent levels of tumor growth inhibition were observed inERþ/HER2�, ERþ/HER2þ as well as biomarker selected TNBCxenografts in response to abemaciclib. In addition, abemaciclibcombined with hormonal blockade and/or HER2-targeted ther-apy induced significantly improved antitumor activity. CDK4/6

inhibition with abemaciclib combined with antimitotic agents,both in vitro and in vivo, did not antagonize the effect of eitheragent. Finally, we identified a set of Rb/E2F-regulated genes thatconsistently track with growth inhibitory response and constitutepotential pharmacodynamic biomarkers of response to abe-maciclib. Taken together, these data represent a comprehensiveanalysis of the preclinical activity of abemaciclib, used alone orin combination, in human breast cancer models. The subtypesmost likely to respond to abemaciclib-based therapies can beidentified by measurement of a specific set of biomarkersassociated with increased dependency on cyclinD:CDK4/6:Rbsignaling. These data support the clinical development ofabemaciclib as monotherapy or as a combination partner inselected ERþ/HER2�, HER2þ/ERþ, and TNBCs. Mol Cancer Ther;17(5); 897–907. �2018 AACR.

IntroductionAdvances in our understanding of the molecular diversity of

human breast cancer have led to the development of specific,molecularly targeted therapies with subsequent improvement inclinical outcomes for patients with hormone receptor positive(ERþ) and human epidermal growth factor receptor-2 positive(HER2þ) disease (1–5). Despite these advances, a number ofpatients with these subtypes who receive appropriate therapiesdemonstrate either de novo or acquired resistance, particularly inthe metastatic setting (6, 7). Investigations into the mechanismsunderlying therapeutic resistance have focused on dysregulationof key signaling pathways downstream of ER and/or HER2 altera-tions, including alterations in the PI3K/AKT/mTOR pathway andmore recently the cyclin D:CDK4/6:Rb:p16 axis (8–12).

The cyclin-dependent kinases 4 and 6 (CDK4/6) are key reg-ulators of the G1–S-phase restriction checkpoint and pharmaco-logically targeting these proteins in combination with hormonalblockade has been shown to provide significant therapeuticbenefit to patients with advanced ERþ/HER2� breast cancer(13–15). In ER-dependent tumors, activated ER signaling leadsto increased transcription and synthesis of cyclin D1, promotingformation of activating complexes with CDK4/6 (16). Phosphor-ylation of the retinoblastoma protein (Rb) by the activatedCDK4/6:cyclin D1 complex removes the sequestration of E2Fby Rb, allowing transcription of factors that promote progressionof the cell cycle through the G1–S-phase restriction point (17).

Mitogen-activated CDK4/6:Rb signaling can also occur inde-pendent of estradiol-mediated ER signaling. It is known thatactivatedHER2 signaling leads to increased transcription of cyclinD1 via activation of the PI3K andMAPK signaling pathways (18).Mitogenic control of the CDK4/6:Rb axis is frequently lostthrough a series of molecular alterations known to occur to agreater or lesser degree, in all subtypes of breast cancer (19, 20).These alterations are often complex and multifactorial, making itdifficult to predictwhichdisease subtypeswill respond toCDK4/6inhibition in the absence of biologic data.

Prior preclinical data from our laboratory demonstrated thata majority of luminal ERþ breast cancer cell lines are sensitiveboth in vitro and in vivo to the CDK4/6 inhibitor, palbociclib, andthat a combination of this drug with hormonal blockade istherapeutically synergistic in this subtype. These data led to theclinical development and recent approval of palbociclib for use in

1Division of Hematology/Oncology, Department of Medicine, Geffen School ofMedicine at UCLA, Los Angeles, California. 2Oncology Discovery Research, LillyResearch Laboratories, Indianapolis, Indiana.

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

Corresponding Authors: Dennis J. Slamon, Geffen School of Medicine at UCLA,10945 Le Conte Avenue, 3360 PVUB, Los Angeles, CA 90095-3152. Phone: 310-825-5193; Fax: 1-310-825-6192; E-mail: [email protected]; and NeilO'Brien, E-mail: [email protected]

doi: 10.1158/1535-7163.MCT-17-0290

�2018 American Association for Cancer Research.

MolecularCancerTherapeutics

www.aacrjournals.org 897

on February 15, 2021. © 2018 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst February 26, 2018; DOI: 10.1158/1535-7163.MCT-17-0290

http://crossmark.crossref.org/dialog/?doi=10.1158/1535-7163.MCT-17-0290&domain=pdf&date_stamp=2018-4-10

http://mct.aacrjournals.org/

ERþ/HER2� metastatic breast cancer (13, 21–23) and subse-quently, a second CDK4/6 inhibitor, ribociclib in the same breastcancer subtype (15). Abemaciclib (LY2835219; Eli Lilly) is also anorally available ATP-competitive inhibitor of CDK4 and CDK6,with single-digit nanomolar potency against both kinases. Basedon early monotherapy clinical data (24) abemaciclib has beengiven "Breakthrough Therapy" status from the FDA. This drug hasnow completed phase II and III clinical development in advancedERþ/HER2�breast cancer resulting in its recent regulatory approv-al for ERþ/HER2� breast cancer either alone or in combinationwith hormonal blockade (25, 26).

In the current study, we wanted to further evaluate the preclin-ical activity of abemaciclib in cell lines representing each of themajor therapeutic molecular subtypes of breast cancer and com-pare it with other approved drugs in this class. Abemaciclib isknown to have increased selectivity for CDK4 over CDK6 as wellas some inhibitory activity at other kinases such as CDK9, PIM1,HIPK2, andDYRK2 (27). In addition, we performed genomic andproteomic biomarker analyses in order to identify biomarkersand/or molecular profiles that correlate with sensitivity or resis-tance to this molecule. Finally, we interrogated the potential fortherapeutically beneficial dual treatment regimens combiningabemaciclib with cytotoxic therapies that are currently approvedand/or commonly used in the management of breast cancer.These data provide insight into biomarkers of response, othermolecular subtypes of breast cancer that may respond to CDK4/6therapy and additional potential combination strategies that maybe appropriate for clinical development.

Materials and MethodsCell lines, cell culture, and reagents

The growth inhibitory activity of abemaciclib (LY2835219-mesylate salt, provided by Richard Beckmann at Eli Lilly), palbo-ciclib (PD-0332991-HCL; Selleck Chemicals) and ribociclib(NVP-LEE011-succinate; MedChem Express) were assessed usinga panel of 44 molecularly characterized human breast cancer celllines representing the known therapeutic subtypes of the disease(22). Cells were cultured in appropriate culture media (e.g.,RPMI 1640, DMEM, L-15) supplemented with 10% to 15%heat-inactivated fetal bovine serum (FBS), 2 mmol/L glutamine,and 1% penicillin G-streptomycin-fungizone solution (PSF,Irvine Scientific) as previously described (28).Cellswere routinelyassessed for mycoplasma contamination using a multiplex PCRmethod and STR profiling by the GenePrint 10 System (Promega)was used for cell-line authentication. Trastuzumab-resistant(BT-474-TR) cells were established as previously described (12, 29).

In vitro proliferation assaysCellswere seeded into 24-well plates in the relevant culturemedia

in duplicate at 5,000 to 50,000 cells perwell, as describedpreviously(22) and the following day 10 mmol/L of abemaciclib, palbociclib,or ribociclib with 2-fold serial dilutions over 6 to 12 concentrationswas added to generate dose–response curves for IC50 determination(Supplementary Procedures). Drug combination studies using doc-etaxel (Taxotere, Hospira) and Carboplatin (Hospira) were per-formed as outlined in the Supplementary Procedures.

Molecular characterization of the breast cancer cell panelTo identify potential predictive mutational biomarkers, whole

exome sequencing was performed using Agilent SureSelect hybrid

capture technology paired with Illumina HiSeq sequencingaccording with manufacturer's protocols (Supplementary TableS1). The deviations from consensus "normal" sequences in thesedata were further filtered to enrich for mutations that are mostlikely to be somatic alterations with functional consequence incancer using multiple published resources, most notably theCOSMIC catalog of somatic mutations in cancer (http://www.sanger.ac.uk/science/tools/cosmic). DNA copy-number altera-tions were determined using comparative genomic hybridizationmicroarray assays (a-CGH)with theAgilent 105KoligonucleotideCGHchip accordingwithmanufacturer's protocols. Knownonco-genes with log2 ratios >1 (2-fold) were considered amplified andknown tumor suppressor genes log2 ratios <0.8 were consideredhomozygous deletions. Baseline total and phosphoprotein levelsof a list of >280 protein analytes enriched for proteins currentlyknown to be involved in cancer biology, were determined usingreverse phase protein array (RPPA) from the core service at theMDAnderson Cancer Center. Cell preparation and analysis wereperformed in accordancewithMDAndersonpublishedprotocols.Data are presented here as log2 (intensity) values. Molecularmarkers commonly tested in breast cancer, such as ER and HER2,and several biomarkers that have previously beenhypothesized toplay a role in response to CDK4/6 inhibition were examined forassociation with response to abemaciclib (listed in Supplemen-tary Table S2).

In vivo efficacy studiesXenograft models of ERþ/HER2�, HER2þ/ERþ, and TNBC

breast cancer cell lines were established in 6-week-old CD-1athymic nude mice (Charles River Laboratories) as described inSupplementary Materials. For ERþ/HER2� and HER2þ/ER� stud-ies, 17-ß-estradiol 60-day release pellets (Innovative Researchof America) were implanted subcutaneously into the left flank7 days before tumor inoculation. For in vivo studies, Fulvestrant(Faslodex, AstraZenica), trastuzumab (Herceptin, Genentech),and docetaxel (Taxotere, Hospira)were purchased from theUCLApharmacy and 4-hydroxytamoxifen was purchased from Sigma-Aldrich. Statistical differences between treatment arms at specifictime points were performed using a two-tailed paired Student ttest. Differences between groups were considered statisticallysignificant at P<0.05. All statisticswere calculated usingMicrosoftExcel. All animal work was carried out under a protocol approvedby IACUC and the UCLA Animal Research Committee.

siRNA knockdown of RB1MDA-231 cells were seeded in 24-well plates at 10,000 cells per

well, followed by transfection with 1.25 pmol siRNA-targetingexon 19 of RB1 gene (Thermo Fisher Scientific, cat#4390824,s522), and 1.5 mL of lipofectamine (Thermo Fisher Scientific) inOpti-MEM I Reduced Serum Medium. Silencer Select NegativeControl No.1 siRNA (Thermo Fisher Scientific, cat#4390843) wasused as a control in all assays. siRNA knockdown was confirmedby Western blot.

Modaplex RT-PCR assayTotal RNA was isolated from homogenized snap-frozen tumor

tissue using the RNeasy kit from Qiagen. Specific cDNA wasproduced from the RNA using SuperScript VILO MasterMix asdescribed by the User's Guide. mRNA expression of a set of cell-cycle targets was quantified using the MODAplex RNA Arraymultiplex qPCR platform (Qiagen). Data were exported from the

O'Brien et al.

Mol Cancer Ther; 17(5) May 2018 Molecular Cancer Therapeutics898



http://www.sanger.ac.uk/science/tools/cosmic

http://www.sanger.ac.uk/science/tools/cosmic


MODAplex software and analyzed in Microsoft Excel. Meanexpression values for each transcript were generated from at leastthree replicate tumor samples. Expression values for each targettranscript are represented as a ratio of the mean expression in theexperimental arm relative to the vehicle control armof the study. A"dose-dependent" response was defined as a >10% inhibition atlow dose (50 mg/kg) abemaciclib accompanied by a furtherincrease in inhibition of >10% at higher dose (75 mg/kg) ofabemaciclib.

ResultsActivity of abemaciclib in the breast cancer cell line panel

The antiproliferative activity of abemaciclib was assessed in apanel of 44 breast cancer cell lines, representing the knownhistological subtypes of the disease (see Supplementary TableS2 for full details). The cell lines were differentially responsive toabemaciclib over a wide concentration range, with IC50 valuesvarying between 0.012 and 3.73 mmol/L (Fig. 1A). On-mecha-nism activity of abemaciclib was confirmed by a loss of Rbphosphorylation in response to treatment (Fig. 1B) followed byinduction of G1 cell-cycle arrest (Supplementary Fig. S1). No lossof pRb or induction of G1 cell-cycle arrest was observed in the Rb-deficient MDA-468 cell line (Fig. 1B; Supplementary Fig. S1).

ERþ/HER2� and HER2-amplified breast cancer cell lines thatexpress high levels (>median) of ER protein (i.e., HER2þ/ERþ;Fig. 1A; Supplementary Table S2) were among the most sensi-tive to abemaciclib. In addition, despite relatively low levels ofER and HER2, a subset of TNBC cell lines also responded to

abemaciclib at IC50 values below500nmol/L (Fig. 1A; Table 1). Ingeneral, sensitivity to abemaciclib was associated with high ERlevels, Rb wild-type status, high Rb total, and phosphoproteinlevels, lack of cyclin E amplification (normal copy number), lowcyclin E protein, p16 deletion, and/or low p16 protein levels(Supplementary Table S1).

In the same panel of cell lines, strong correlations were iden-tified between response to abemaciclib and two other CDK4/6inhibitors, palbociclib and ribociclib (Fig. 1C–E). Hormonereceptor–positive cell lines (either HER2 normal or amplified)were commonly sensitive to all three molecules. In comparativeanalyses, abemaciclib was the most potent of the three moleculestested, followed by palbociclib and then ribociclib. In biomarker-positive breast cancer cell lines, the average (geometric mean)IC50 of abemaciclib was 168 nmol/L, compared with 306 nmol/Lfor palbociclib and 913 nmol/L for ribociclib.

Abemaciclib in combination with hormone blockade inERþ/HER2� breast cancer cell xenografts

Combined activity of abemaciclib and hormone blockade wasconfirmed in twoERþbreast cancer cell lines xenografts (Fig. 2 andSupplementary Figure S2). Significant tumor regressions wereobserved with single-agent abemaciclib in MCF-7 (ERþ/HER2�)xenografts, and combination with either tamoxifen or fulvestrantinduced a marginal increased benefit in antitumor response(Fig. 2A and B; Supplementary Table S3). However, within the1st week of drug withdrawal, tumors began to progress in themice treated with 50 mg/kg of abemaciclib alone as well inthe mice treated with either single-agent tamoxifen or fulvestrant.

R² = 0.6447

1.00E-02

1.00E-01

1.00E+00

1.00E+01

1.00E-02 1.00E-01 1.00E+00 1.00E+01

A BPa

lboc

iclib

Log

IC50

μmol

/L

Abemaciclib Log IC50 μmol/L

Ribo

cicl

ibLo

g IC

50μm

ol/L

Abemaciclib Log IC50 μmol/L

Ribo

cicl

ibLo

g IC

50μm

ol/L

Palbociclib Log IC50 μmol/L

Log

IC50

μmol

/L

HER2-/ER+

HER2+/ER+ TNBC

HER2+/ER-

DC E

EFM-19 ZR-75-1 MDA-361 MDA-468Abema 200 nmol/L

pRbS807/811

Total-Rb

Cyclin D1

CK-19

- + - + - + - +

b-Ac�n

R² = 0.5956

1.00E-02

1.00E-01

1.00E+00

1.00E+01

1.00E-02 1.00E-01 1.00E+00 1.00E+01

R² = 0.7606

1.00E-02

1.00E-01

1.00E+00

1.00E+01

1.00E-02 1.00E-01 1.00E+00 1.00E+01

0.001

0.01

0.1

1

10

MDA

-MB-

134-

VIM

DA-M

B-45

3M

DA-M

B-17

5-VI

IM

DA-M

B-41

5T-

47D

HCC1

500

ZR-7

5-30

CAM

A-1

UAC

C-73

2EF

M-1

9SU

M-1

90PT

MCF

-7HC

C141

9U

ACC-

812

MDA

-MB-

361

ZR-7

5-1

HCC2

218

HCC2

02BT

-474

HCC1

395

BT-2

0EF

M-1

92A

CAL-

51M

DA-M

B-23

1SK

-BR-

3HC

C114

3SU

M-2

25CW

NCO

LO-8

24HC

C180

6HC

C195

4Hs

578T

HCC7

0U

ACC-

893

MDA

-MB-

436

MDA

-MB-

468

HCC1

937

JIMT-

1M

DA-M

B-15

7HC

C156

9HC

C38

HCC1

187

DU44

75BT

-549

Figure 1.

Activity of abemaciclib in breast cancer cell lines. A, In vitro IC50 values (generational inhibition) for each of the breast cell lines. Data, mean IC50 � 95%confidence interval where available. B, Effect of single-dose (200 nmol/L) abemaciclib treatment on Rb signaling, cell lysates were prepared 5 days postdosing.C, Scattergram comparing the in vitro IC50 values of abemaciclib with palbociclib in matching breast cancer cell lines; D, abemaciclib versus ribociclib; E,palbociclib versus ribociclib, for scattergrams, cell line subtypes are color coded consistent with A. R values were calculated in Excel. All experiments wererepeated in at least duplicate.

Preclinical Activity of Abemaciclib in Breast Cancer

www.aacrjournals.org Mol Cancer Ther; 17(5) May 2018 899




Conversely, sustained inhibition of tumor growth was observedfor over 6 weeks after drug withdrawal in the majority of animalstreated with either drug combination or with high-dose abema-ciclib monotherapy at 75 mg/kg (Fig. 2C). Molecular analysis ofxenograft tissue at day 4 of dosing revealed a greater loss of pRb

in response to the combination versus single-agent treatments(Fig. 2D). This was accompanied by a reduction in mitosis asindicated by a decrease in FOXM1 protein levels (Fig. 2D).Western blot analysis of residual xenograft tissue collected at theend of a recovery phase of 6 to 7 weeks revealed that active cell

BA

0

200

400

600

800

1,000

1,200

0 5 10 15 20 25 30 35Vehicle Abema 50 mg/kg Abema 75 mg/kgFulvestrant Abema+Fulv TamoxifenAbema+Tam

Tum

or v

olum

e (m

m3 )

*

MCF-7: ER+ BCL

C

Tum

or v

olum

e (m

m3 )

0

400

800

1,200

1,600

0 15 30 45 60 75 90

Abema 50 mg/kg

Treatment withdrawn

0

400

800

1,200

1,600

0 15 30 45 60 75 90

Abema 75 mg/kg

Treatment withdrawn

0

400

800

1,200

1,600

0 15 30 45 60 75 90

Tamoxifen

Treatment withdrawn

0

400

800

1,200

1,600

0 15 30 45 60 75 90

Fulvestrant

Treatment withdrawn

0

400

800

1,200

1,600

0 15 30 45 60 75 90

Abema + Tamox

Treatment withdrawn

0

400

800

1,200

1,600

0 15 30 45 60 75 90

Abema + Fulv

Treatment withdrawnDays

E > 6 weeks post drug withdrawal

1N 1R 1L 1B 2L 2N 2B 4R 4B 4N 5R 5B 5L 6L 6R 6N 7R 7L 7N

Vehicle Abema Fulvestrant Abema + fulv Tamox Abema + tam

pRbS807/811

Total-Rb

FOXM1

Kera�n-19

Vehicle Abema Fulvestrant Abe + fulv Tamox Abe + tam

D1R 1B 2L 2R 4N 4B 4N 5N 5R 6B 6N 7N 7B

pRbS807/811

Total-Rb

FOXM1

b-Ac�n b-Ac�n

Cyclin-D

On treatment (Day 3)

DAY 6: 200x

Vehicle Abema Fulvestrant

Tamoxifen Abema+Fulvestrant Abema+Tamoxifen

F

-150

-100

-50

0

50

100

150

200

250

300

ΔT/C

and

Reg

ress

ions

(%)

Vehicle Abe 50mg/kg

Abe75mg/kg

Fulv Abe +

Fulv

Abe +

Tam

Tam

Progressive disease Stable disease Par�al responseComplete response

Figure 2.

Induction of sustained tumor growth inhibition by abemaciclib plus hormone blockade. A, Growth curves for MCF-7 (ERþ) breast cancer cell line (BCL) xenograftstreated with either single abemaciclib or combination with hormonal blockade (fulvestrant or tamoxifen), 8 mice per arm. For combination arms, 50 mg/kgabemaciclib was used. Data, mean tumor volume � SEM. B, Waterfall plots representing the % progression or tumor regression in each individual tumorwithin each treatment arm, �20% progression was considered stable disease, �50% tumor regression was considered a partial response, and 100% regressiona complete response. Fulv, Fulvestrant; Tam, Tamoxifen. C, Growth curves for individual xenograft tumors posttreatment withdrawal relative to mean tumorvolumes for the vehicle control.D,Western blot analysis of snap-frozen tumor tissue taken frommice 24 hours after the 3 days of continuous dosing. E,Western blotof representative tumor lysates taken >6 weeks postdrug withdrawal. F, Representative images of cells stained for b-galactosidase 6 days posttreatmentwith the molecules indicated. � , Statistically significant difference from vehicle control; P < 0.05.

Table 1. Biomarkers of response to abemaciclib triple-negative breast cancer cell lines

Cell lineIC50

mmol/L ER HER2PIK3CAPMs

RB1PMs

RB1CN Rb RbpS807S811

CDKN2APMs

CDKN2ACN p16

CCNE1CN CyclinE1

HCC1395 0.310 -2.16 -0.1 Wt Wt GAIN 0.44 3.15 Wt HD -3.01 GAIN 0.27BT-20 0.328 -1.61 1.52 p.Pro539Arg Wt NC 0.56 2.98 Wt HD -2.81 NC -0.33

CAL-51 0.407 -2.1 1.62 p.Glu542Lys Wt NC 0.89 3.84 Wt NC -2.15 NC 0.07

MDA-MB-231 0.443 -2.26 0.79 Wt Wt NC 0.65 3.43 Wt HD -2.73 GAIN -0.17

HCC1143 0.528 -1.15 0.97 Wt Wt NC 0.36 2.87 Wt LOH -2.55 NC -0.17

COLO-824 0.903 -1.7 1.56 Wt Wt HD -0.31 1.15 Wt GAIN -0.3 NC 0.48

HCC1806 0.904 -1.76 0.68 Wt Wt NC 0.35 3.07 Wt HD -2.96 HIGH AMP 1.98

Hs578T 0.968 -2.18 0.86 Wt Wt NC 0.74 3.56 Wt NC -1.39 NC -0.31

HCC70 0.971 -1.96 1.74 Wt Wt NC -0.08 0.8 Wt AMP -0.31 GAIN 0.21

MDA-MB-436 1.050 -1.98 0.66 Wt Wt GAIN -0.26 0.24 Wt NC -0.69 NC 0.83

MDA-MB-468 1.090 -1.95 -0.53 Wt Wt HD -0.41 1.52 Wt NC -1.02 AMP 0.27

HCC1937 1.180 -1.22 1.08 Wt c.2221*>-G LOH -0.32 1.13 Wt NC -0.51 NC 0.09

MDA-MB-157 1.520 -2.1 0.63 Wt Wt LOH 0.11 1.87 Wt NC -0.23 AMP 1.02

HCC38 2.280 -2.47 2.42 Wt Wt NC 1.48 3.78 Wt NC -3.29 NC -0.58

HCC1187 2.300 -1.92 2.37 Wt Wt NC -0.28 1.37 Wt NC -0.03 GAIN 1.09

DU4475 3.290 -2.17 -0.4 Wt Wt HD 0.28 3.4 Wt NC -2.21 NC 1.54

BT-549 3.730 -2.36 0.91 Wt Wt HD -0.54 0.43 Wt GAIN 0.64 NC 0.23

Abbreviations: CN, copy number; PM, point mutation.

O'Brien et al.





division and Rb signaling were restored in the 50 mg/kg abema-ciclib and single-agent endocrine treatments, as indicated by therecovery of FOXM1 and pRb protein levels (Fig. 2E). In contrast,FOXM1/pRb signal remained low in xenograft tissues collectedfrom combination arms at the same time point (Fig. 2E). Toinvestigate if the sustained reduction in Rb/FOXM1 signal in thecombination arms was due to induction of permanent growtharrest/senescence or induction of cell death, we measured levels ofthe human epithelial marker protein keratin-19 (CK-19), a struc-tural protein not directly regulated by abemaciclib treatment(Supplementary Fig. S3). In the MCF-7 xenograft tissues, CK-19protein levels were significantly reduced or lost in the combinationtreatment arms, particularly in abemaciclib plus tamoxifen treatedmice, indicating loss of human epithelial tumor cells (Fig. 2E).Using a ß-galactosidase–staining assay, in vitro assays of the MCF-7cells showed increased induction of senescence in response to thecombination as opposed to the single agent (Fig. 2F).

Activity of abemaciclib in HER2þ/ERþ breast cancer cell linesand xenografts

Activity of abemaciclib single agent or combination with HER2and ER-targeted therapy was assessed in two xenograft models ofHER2þ/ERþ breast cancer, parental BT474 cells and BT474 cellsconditioned through long-term drug exposure to progress ontrastuzumab therapy (BT-474-TR (12)). Single-agent abemaciclibinduced significant tumor growth inhibition (TGI) in both mod-els (BT-474; P ¼ 0.028, BT-474-TR; P < 0.001; Fig. 3A–C; Sup-plementary Table S4). The combination of abemaciclib withtrastuzumab induced significantly improved (P ¼ 0.0012) TGIs

and tumor regressions in xenografts progressing on trastuzumabalone (Fig. 3B and C; Supplementary Table S4). The triple com-bination of abemaciclib, trastuzumab, and tamoxifen furtherimproved tumor regressions in both models (Fig. 3A–C; Supple-mentary Table S4). In vitro studies confirmed that triple blockadeof CDK4/6, HER2, and ER signaling leads to a greater inductionof cell death in both trastuzumab-sensitive and -resistantHER2þ/ERþ breast cancer cell lines (Fig. 3D). Abemaciclib wasalso shown to combine effectively with docetaxel in these HER2-amplified breast cancer cell line xenografts. The addition ofdocetaxel to abemaciclib did not inhibit the TGI induced byabemaciclib or docetaxel alone. Moreover, the combination ofabemaciclib, trastuzumab, tamoxifen, and docetaxel was themost efficacious arm in both studies (Fig. 3A–C; SupplementaryTable S4). Marginal and reversible body weight loss was observ-ed in the mice treated with docetaxel containing arms over the4 weeks treatment period (Supplementary Table S4).

Western blot analysis of xenograft tissues collected after 4 daysof treatment confirmed a dose-dependent loss of pRb, TOPOIIa,and pHH3 in mice treated with abemaciclib (Fig. 3E). Furtherreduction in Rb signaling was observed in xenografts treated withabemaciclib plus trastuzumab or trastuzumab plus tamoxifen.The addition of docetaxel did not block the pRb knockdowninduced by abemaciclib or the combination of abemaciclib withtrastuzumab plus tamoxifen (Fig. 3E).

Identification of a subset of TNBC cells sensitive to abemaciclibIn TNBC cell lines, where cell growth is independent of ER

status, response to abemaciclib appears to remain dependent on

A BT-474, HER2+/ER+, Trastuzumab sensi�ve

Tum

or v

olum

e (m

m3 )

0

100

200

300

400

500

600

0 4 8 12 16 20 24 28

Vehicle Abe 50 mg/kg TrastuzumabAbe+Tz Abe+Dx Abe+Tz+TamAbe+Tz+Dx+Tam

Days

*

#

Days

B

0

200

400

600

800

1,000

0 4 8 12 16 20 24 28

Vehicle Abe 50 mg/kg Abe 75 mg/kgTrastuzumab Abe+Tz Abe+DxAbe+Tz+Tam Abe+Tz+Dx+Tam

BT-474-TR, HER2+/ER+, Trastuzumab acquired resistant

Tum

or v

olum

e (m

m3 )

*

#

0

5

10

15

20

25

30

35

BT474 BT-474-TR

Vehicle Abema 200 nmol/L Traz 15 µg/mLTam 500 nmol/L Traz+Tam Abema+TamAbema+Traz Abema+Traz+Tam

% A

nnex

inV +

cells

DVehicle Abe (50) Abe (75) Trast (Tz) Abe (50)+Tz

1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 4

pRbS807/811

TOPO IIα

PHH3

Total HER2

b-Ac�n

Abe+Dtx

1 2 3 4

Abe+Tz+Tam Abe+Tz+Tam+Dtx

1 2 3 4 1 2 3 4

E

C

-150

-50

50

150

250

350

450

550

BT-474

BT-474-TR

Δin

Tum

or v

olum

e (m

m3 )

Figure 3.

Activity of abemaciclib in combination with HER2 and ER-targeted therapies in HER2þ/ERþ breast cancer. A and B, Growth curves for BT-474 and trastuzumabconditioned BT-474-TR cell line xenografts treated with either single agent abemaciclib ("Abe") or combination with HER2-targeting (trastuzumab, "Tz"),ER-targeting (tamoxifen, "tam") or SOC chemotherapy (docetaxel, "Dtx"), 8 mice per arm. Fifty mg/kg abemaciclib was used in combination arms. C, Changein tumor volume from baseline for each experimental arm. D, Induction of apoptosis in cell line in 2D cultures treated with combination CDK4/6, HER2/ERcombination therapy. Histograms, % annexin V-positive (AN Vþ) cells after 5 days of treatment. Data, mean � SD. All experiments were repeated in at leastduplicate. E, Lysates from snap-frozen BT474-TR tumor tissue, collected after 4 days of dosing, were subject to immunoblotting with the antibodies indicated.� , Statistically significant difference compared with vehicle control; #, statistically significant difference compared with single-agent trastuzumab; P < 0.05.






intact Rb signaling. Those TNBC cell lines that have highbaseline levels of total and phosphorylated pRb, accompaniedby low levels of p16 protein, are among the most sensitive toabemaciclib (Table 1). Cell lines with copy-number gains oramplification of the CCNE1 (cyclin E1) gene are clearly lessresponsive (Table 1).

Xenografts of TNBC cell lines expressing sensitive and resis-tant biomarker profiles were measured for response to abema-ciclib single agent and combination with docetaxel, a standard-of-care agent used for TNBC. Consistent with all the precedingdata, induction of tumor regressions or stable disease wasobserved only in the MDA-231 (Fig. 4A; Supplementary TableS5) and BT-20 xenografts (Supplementary Fig. S4) that havehigh levels of pRb and low levels of p16 at baseline (Fig. 4C.left). Conversely, HCC70 xenografts with low levels of pRb andhigh p16, continued to progress through abemaciclib mono-therapy (Fig. 4B). On-target activity of abemaciclib was con-firmed in the sensitive MDA-231 xenografts by reduction intotal and phosphorylated Rb and the cell-cycle progressionmarker, FOXM1, in response to treatment (Fig. 4C, right).Consistent with our observations in HER2þ/ERþ breast cancerxenografts, combination of abemaciclib with docetaxel did notantagonize the activity of either single agent (Fig. 4A and B).Finally, the role of pRb in predicting response to abemaciclib inTNBC was confirmed by siRNA knockdown of RB1 gene expres-sion in the MDA-231 cells. Subsequent reduction of baselinepRb levels significantly reduced the sensitivity of the MDA-231cells to abemaciclib (Fig. 4D).

Combined activity of abemaciclib and cytotoxicchemotherapy

The potential of abemaciclib to be used in combination withcytotoxic chemotherapy was further investigated in a cell linemodel of TNBC in vitro. In order to determine if the lack ofantagonism observed in our in vivo models could be attributedto an effect of the timing of drug administration, we investigateddosing strategies comparing coadministration versus sequencingof these treatments.

Simultaneous treatment of MDA-231 cells with abemaciclibplus docetaxel or carboplatin resulted in increased inhibitionof cell proliferation compared with single-agent treatments(Fig. 5A). Concentrations of abemaciclib below 10 nmol/Linduced profound (>60%) inhibition of cell proliferation whencombined with either antimitotic agent. Pretreatment of cellswith abemaciclib for 2 days reduced cell proliferation rate viainduction of G1 cell-cycle arrest (Supplementary Fig. S5A)without impacting the activity of docetaxel or carboplatinrelative to cells pretreated with vehicle control (Fig. 5B andC). Pretreatment with abemaciclib for 24 hours also inducedG1 cell-cycle arrest (Supplement Fig. S5B) without blockingapoptosis induced by high-dose docetaxel treatment. In con-trast, a dose-dependent decrease in apoptosis induction bydocetaxel was observed in response to pretreatment with palbo-ciclib or ribociclib (Fig. 5D). Coadministration of docetaxelwith either palbociclib or ribociclib after CDK4/6-inhibitorpretreatment blocked apoptosis induction, whereas cotreat-ment with abemaciclib did not (Fig. 5E).

A

B

C

D0

100

200

300

400

500

0 3 6 9 12 15 18 21

Vehicle control Abemaciclib 50 mg/kg QDAbemaciclib 75 mg/kg QD Docetaxel 10 mg/kg QWAbemaciclib + Docetaxel

Tum

or v

olum

e (m

m3 )

ΔT/C

and

Reg

ress

ions

(%) Vehicle Abe

50 mg/kgAbe

75 mg/kgDocetaxel Abema

+Dtxl

-100

-50

0

50

100

150

200

Progressive disease Stable disease Par�al response

Complete response

MDA-231, High pRb, Low p16 MDA-231: Day 21

0

200

400

600

800

1,000

1,200

0 3 6 9 12 15 18 21

Vehicle control Abemaciclib 50 mg/kg QD

Abemaciclib 75 mg/kg QD Docetaxel 10 mg/kg QW

Abemaciclib + Docetaxel

Tum

or v

olum

e (m

m3 )

HCC70, Low pRb, High p16

-100

-50

0

50

100

150ΔT

/C a

nd R

egre

ssio

ns (%

) Vehicle Abe 50 mg/kg

Abe75 mg/kg

Docetaxel Abema+

Dtxl

Progressive disease Par�al response

1N 1R 1L 1N

pRbS807/811

HCC70 MDA231

p16

b-Ac�n

pRbS807/811

b-Ac�n

FOXM1

Cyclin D

1L 1N 2L 2R 3L 3N

Vehicle Abema 50 Abema 75

Total-Rb

76.4

49.4

23.9

75.8

43.5

20.6

46.4

23.7

15.0

0

10

20

30

40

50

60

70

80

90

100

1,000 nmol/L Abema 200 nmol/L Abema 40 nmol/L Abema

Untransfected

Scrambled control

Rb1 Knockdown

% G

ener

a�on

al in

hibi

�on

pRbS807/811

tRb

b-Ac�n

Abema 200 nmol/L - - - - - - - + + + - - - + + +

Day 0 Day 2 Day 5

Figure 4.

Abemaciclib has activity in xenograft models of TNBC. A and B, Growth curves for TNBC cell line xenografts treated with either single abemaciclib ("Abe")or combination with SOC chemotherapy (docetaxel "dtxl"), 8 mice per arm. Fifty mg/kg abemaciclib was used in combination arms. C, Left, Western blotanalysis of snap-frozen tumor tissue taken from baseline vehicle control arms of each xenograft study. Right, Western blot analysis of snap-frozen tumor tissue,collected from mice in each arm of the MDA-231 study, 24 hours after the 3 days of continuous dosing. D, siRNA knockdown of Rb1 confirms Rb1 signaling isrequired for abemaciclib response. Histogram, % generational inhibition in response to 5 days abemaciclib treatment in MDA-231 with and without Rb1 siRNAknockdown. Western blots confirm knockdown of Rb1 by siRNA at the time points indicated. #, Statistically significant versus scrambled nontargetingcontrol, P value according to Student t test.

O'Brien et al.





Identification of a pharmacodynamic signature of response toabemaciclib treatment

To identify potential pharmacodynamic (PD) markers ofresponse to abemaciclib, we used a Modaplex PCR-based plat-form tomeasure changes in the expressionof a set of 22 cell-cycle–associated genes in pre- and posttreatment (day 4) samples fromtumor tissues responding to abemaciclib-based therapy. InMCF-7ERþ/HER2� xenografts, a rangeofmRNAexpression changeswereinduced by treatment with abemaciclib, hormonal blockade or acombination of the two (Fig. 6A). A subset of 10 transcriptsshowed a dose-dependent reduction in expression in responseto low- and high-dose abemaciclib. Single-agent treatment witheither fulvestrant or tamoxifenhad little effect on the expressionofthese 10 candidate PD biomarkers. However, combination oflow-dose abemaciclib (50 mg/kg) with either fulvestrant ortamoxifen increased the inhibition of the candidate gene setcompared with inhibition induced by high-dose abemaciclib(75 mg/kg; Fig. 6B). Using the same criteria, an abemaciclibdose-dependent gene signature was identified from ERþ/HER2þ

xenografts. Similar to the ERþ/HER2�models, expression of thesetranscripts was similarly regulated by CDK4/6-inhibition in com-bination with either trastuzumab or tamoxifen or both (Fig. 6C).Transcripts forRRM2, TOPO2A,MKI67,MCM7,CDK2, andCDK4were all found tobe regulatedby abemaciclib in adose-dependentmanner in sensitive tumors from MCF-7 (ERþ/HER2�), ZR-75-1

(ERþ/HER2�), and BT-474-TR (HER2þ/ERþ) xenografts(Fig. 6D). Expression of five of these transcripts, RRM2, TOPO2A,MKI67, MCM7, and CDK2 are directly regulated by the E2F-transcription factor immediately downstream of Rb, indicatinga direct on-target regulation of these genes by CDK4/6 inhibitionusing abemaciclib.

DiscussionThe cyclinD:CDK4/6:Rb:p16 signaling axis is frequently dysre-

gulated in cancer. Efforts to pharmacologically target this pathwayhave been validated by the approval of the specific CDK4/6inhibitor, palbociclib (PD-0332991, Pfizer), ribociclib (LEE-011,Novartis), and most recently abemaciclib (Eli Lilly), in combi-nation with endocrine-based therapies in advanced ERþ/HER2�

breast cancer (13, 15, 21–23, 25, 26). Abemaciclib is structurallyand biologically distinct from palbociclib and ribociclib withgreater selectivity for CDK4 over CDK6 (27) and clinically,abemaciclib appears to have superior single-agent activity whencompared with the other approved CDK4/6 inhibitors (24, 25,30, 31). It is possible that the CDK4 selectivity of abemaciclibcontributes to the improvement in activity by reducing the degreeof neutropenia usually associated with inhibition of CDK6,which in turn avoids the need for an interruption in dosingassociated with both palbociclib and ribociclib. This difference

0

2,000

4,000

6,000

8,000

10,000

12,000

Veh/

Veh

Veh/

Carb

Veh/

Abe

Veh/

Abe+

Carb

Abe/

Veh

Abe/

Carb

Abe/

Abe

Abe/

Abe+

Carb

Growth Day 2 to Day 5


Cell

num

ber

C

D

Prop

idiu

mio

dide

Annexin V FITC

Vehicle/Vehicle Vehicle/Dtx 100 nmol/L

Abema/Dtx 100 nmol/L Palbo/Dtx 100 nmol/L Ribo/Dtx 100 nmol/L

B

Cell

num

ber

0

2,000

4,000

6,000

8,000

10,000

12,000

Veh/

Veh

Veh/

Dtx

Veh/

Abe

Veh/

Abe+

Dtx

Abe/

Veh

Abe/

Dtx

Abe/

Abe

Abe/

Abe+

Dtx



A

0

20

40

60

80

100

120 Abemaciclib (Abe)Docetaxel (Dtx)Carbopla�n (Carb)Abe+DtxAbe+Carb

%Ge

nera

�ona

linh

ibi�

on

Abema (nmol/L) 1 0.2 0.04 0.008Dtx (nmol/L) 10 2.0 0.40 0.080Carb ( mol/L) 25 12.5 6.25 3.125

* * *

02468

101214161820

Veh/

Veh

Veh/

Dtx

100

nmol

/L

Abe

200

nmol

/L/D

tx

Abe

500

nmol

/L/D

tx

Abe

1,00

0 nm

ol/L

/Dtx

Plb

200

nmol

/L/D

tx

Plb

500

nmol

/L/D

tx

Plb

1,00

0 nm

ol/L

/Dtx

Rib

200

nmol

/L, D

tx

Rib

500

nmol

/L, D

tx

Rib

1,00

0 nm

ol/L

, Dtx

%An

nexi

nV

+ce

lls#

##

#

0

2

4

6

8

10

12

14

16

18

Veh/

Veh

Veh/

Dtx1

00 n

mol

/L

Abe

200

nmol

/L/A

be+D

tx

Abe

500

nmol

/L/A

be+D

tx

Abe

1,00

0 nm

ol/L

/Abe

+Dtx

Plb

200

nmol

/L/P

lb+D

tx

Plb

500

nmol

/L/P

lb+D

tx

Plb

1,00

0 nm

ol/L

/Plb

+Dtx

Rib

200

nmol

/L,R

ib+D

tx

Rib

500

nmol

/L,R

ib+D

tx

Rib

1,00

0 nm

ol/L

,Rib

+Dtx

E

%An

nexi

nV

+ce

lls #

#

#

#

#

0.0016 0.000320.0160 0.00320

1.5635 0.78125

Figure 5.

Abemaciclib combined with cytotoxic chemotherapy induces increased inhibition of cell proliferation and induction of cell death. A, MDA-231 TNBC cells weretreated simultaneously with a range of concentrations of abemaciclib plus docetaxel or carboplatin for 5 days. B and C, Cells were pretreated for 48 hourswith or without 200 nmol/L abemaciclib followed by treatmentwith 1 nmol/L docetaxel (dtx) or 25mmol/L carboplatin (carb) or the combination of abemaciclib plusdocetaxel or carboplatin for the remaining 72 hours. Total cell counts were measured at day 0, day 2 (washout), and day 5. D, Annexin V staining of cellspretreated with or without increasing concentrations of abemaciclib, palbociclib, or ribociclib for 24 hours followed by 24 hours of treatment with 100 nmol/Ldocetaxel, left panels show representative examples, summary histogram on the right. E, Annexin V staining of cells pretreated for 24 hours with a range ofconcentrations of either abemaciclib, palbociclib, or ribociclib followed by switch to 100 nmol/L docetaxel plus abemaciclib, palbociclib, or ribociclib. Veh, vehiclecontrol/untreated. Data, mean � SD. All experiments were repeated in at least duplicate. � , Statistically significant difference compared with vehicle control;#, statistically significant difference compared with vehicle/docetaxel arm; <0.05, P values according to Student t test.






may improve the chances of forcing tumor cells into permanentgrowth arrest and ultimately senescence (24). In this study, wepresent a comprehensive analysis of the preclinical activity ofabemaciclib using both in vitro and in vivo preclinical modelsspanning the known molecular subtypes of breast cancer. Usingpre- and posttreatment data, we have identified a set of consistentbiomarkers of sensitivity to abemaciclib in ERþ, HER2þ, andTNBC subtypes.

Measurement of the growth inhibitory activity of abemaci-clib across a panel of 44 human breast cancer cell lines iden-tified the ERþ/HER2� subtype as most sensitive to CDK4/6inhibition. Despite the wide therapeutic range of this molecule,each of the 9 cell lines classified as ERþ/HER2� had IC50 valuesbelow 200 nmol/L, making further stratification of responsebiomarkers within this subtype unnecessary. This activity wasconfirmed in two xenograft models of ERþ/HER2� breastcancer by the induction of complete arrest of tumor growthusing abemaciclib monotherapy. Inhibition of tumor prolifer-ation was accompanied by dose-dependent decreases of phos-phorylated Rb and induction of cell-cycle arrest as measured byloss of the cell-cycle progression markers, TOPOIIa (S-phase),phosphohistone-H3 (G2–M), and FOXM1 (cellular senes-cence). At the tested doses, abemaciclib did not inhibitCDK-9 signaling, consistent with previous data reported forthis molecule (27, 32). Comparison of the activity of abema-ciclib, palbociclib, and ribociclib across the panel of 44 breastcancer cell lines identified a strong correlation within andbetween the major histologic subtypes of breast cancer for thisclass of molecule. However, of the three, abemaciclib appearedto be the most potent using our assays.

Our studies also confirm that abemaciclib has additive and/orsynergistic activity in combination with endocrine-based thera-pies in cell line xenograft models of ERþ/HER2� breast cancer.Continuous daily treatment using clinically achievable doses ofabemaciclib (50 mg/kg) in combination with tamoxifen or ful-vestrant was well tolerated and led to downregulation of Rbsignaling and significantly improved antitumor responses com-pared with results with either agent alone. While tumors fromsingle-agent–treated animals began to proliferate within days oftreatment cessation, the majority of xenografts from mice treatedwith combination therapy showed no signs of tumor progression,despite withdrawal of treatment for more than 6 weeks. Xeno-grafts collected during the dosing phase of the combinationstudies indicate an initial induction of cellular senescence asmeasured by loss of FOXM1 and in vitro assays showing a greaterinduction of senescence with the combination treatment. Xeno-graft materials collected at the end of the study indicate a loss ofhuman epithelial tumor cell population. The marked tumorregressions observed in the combination arms provide furtherevidence of induction of tumor cell death induced by this regi-men. The data reported here are consistent with those recentlyreported from the MONARCH-2 clinical trial (NCT02107703),which showed significant improvements in progression-free sur-vival (PFS) and objective response rates (ORR) in patients withadvanced breast cancer treated with abemaciclib plus fulvestrantversus fulvestrant alone (26).

Abemaciclib-sensitive cell lines were also identified within theHER2-amplified subgroup. There is considerable preclinical evi-dence to support targeting CDK4/6 in HER2-amplified breastcancers given that HER2 signaling can drive cell-cycle progression

DMCF-7 ZR-75-1 BT474-TR

50 mg/kg Abemaciclib






CCNB1 74.52 92.05 0.50 13.45 88.73 94.00

RRM2 67.20 96.56 52.28 91.20 31.53 93.37

TOPO2A 61.15 84.21 76.94 94.43 78.73 94.33

MKI67 55.43 88.23 70.66 94.22 74.50 94.50

E2F1 32.82 84.09 0.00 55.28 0.00 65.70

MCM7 29.38 76.18 25.37 74.84 28.30 77.37

CDK4 27.21 38.17 17.10 35.95 26.87 44.57

PTEN 15.86 57.55 31.22 30.10 0.00 0.00

E2F2 14.04 84.61 18.53 64.21 -1.93 14.77

CDK2 11.69 46.86 20.32 57.23 11.03 64.63

RBL1 0.00 13.25 24.16 52.63 1.13 0.000

10

20

30

40

50

60

70

80

90

100

TOPO2A MKI67 FOXM1 RRM2 MCM7 CDK4 RB-1 EX25-26 RB-1 EX17-18 CDK2

50 mpk Abemaciclib

75 mpk Abemaciclib

50-AB + Trastuzumab

50-AB + Tamox

50-AB + Trastuzumab + Tamox

C

% In

hibi

�on

of g

ene

expr

essio

n

0

10

20

30

40

50

60

70

80

90

100

50 mg/kg Abemaciclib 75 mg/kg Abemaciclib Fulvestrant Abemaciclib (50 mg/kg)+ fulvestrant

Tamoxifen Abemaciclib (50 mg/kg)+ Tamox

CCNB1 RRM2 TOPO2A MKI67 E2F4 E2F1 CDKN1A MCM7CDK4 PTEN E2F2 CDK2 RB-1 EX17-18 MYC Cyclin E E2F3RBL1 CCND3 RB-1 EX25-26 CCND1 E2F5 RBL2

A%

Inhi

bi�o

n of

gen

e ex

pres

sion

0

10

20

30

40

50

60

70

80

90

100

CCNB1 RRM2 TOPO2A MKI67 E2F1 MCM7 CDK4 PTEN E2F2 CDK2

50 mg/kg Abemaciclib 75 mg/kg Abemaciclib

Abemaciclib (50 mg/kg) + fulvestrant Abemaciclib (50 mg/kg) + Tamox

B

% In

hibi

�on

of g

ene

expr

essio

n

MCF-7 (ER+) Xenogra�s

BT474-TR (HER2+/ER+) Xenogra�s

MCF-7 (ER+) Xenogra�s

Figure 6.

Identification of PD biomarkers of abemaciclib response in vivo. A, Percent inhibition of mRNA expression of 22 cell-cycle–regulated genes, relative to vehiclecontrol, in each of the treatment arms in tumor collected snap-frozen 24 hours after the 3 days of continuous dosing in the MCF-7 xenograft study. B, Percentinhibition of a subset of genes that show abemaciclib dose dependence and response to abemaciclib plus hormone blockade therapy. C, Abemaciclib dose-dependent genes in HER2þ/ERþ breast cancer cell line xenografts are also responsive to HER2/ER-targeted therapy. D, Subset of genes commonly alteredby abemaciclib treatment across three breast cancer cell line xenograft studies. Values, Mean percent inhibition of gene expression relative to vehicle controlin each study.

O'Brien et al.





through activation of cyclinD1:CDK4/6 signaling (18, 33). CyclinD1 is required for the formation of HER2-initiated tumors inmouse models and pharmacologically targeting CDK4/6 blockstumor formation inmice (34, 35). In addition, preclinical studiesshow that targeting CDK4/6 overcomes resistance toHER2-direct-ed therapy both in vitro and in vivo (9, 36). The data presented hereprovide further insight into the specific subtypes of HER2-ampli-fied breast cancers most likely to benefit from CDK4/6-targetedtherapy. Analysis of abemaciclib response within the panel of 17HER2-amplified breast cancer cell lines indicates that HER2-amplified cell lines with higher levels of ER protein accompaniedby intact downstream Rb signaling (high total Rb and pRb and nocyclin E amplification) are most sensitive to abemaciclib treat-ment. These data indicate that HER2-amplified breast cancer withhigher levels of ER (i.e., HER2þ/ERþ) may be more susceptible toCDK4/6 intervention than those that are HER2þ/ER�. The currentdata demonstrate that single-agent abemaciclib induces signifi-cant tumor growth inhibition as well as dose-dependent inhibi-tion of pRb signaling and cell-cycle arrest in HER2þ/ERþ xeno-grafts. In addition, abemaciclib combined effectively with trastu-zumab and in triple combination with tamoxifen, induced sig-nificant regressions in both trastuzumab-sensitive and -resistantHER2þ/ERþ xenografts. In vitro studies confirmed that the triplecombination of targetingCDK4/6,HER2, andER leads to a greaterinduction of cell death in cells when compared with single-agenttreatment. The specific mechanism by which this triple combi-nation is effective remains to be determined. Finally, abemaciclibactivity in HER2-amplified tumors progressing on trastuzumabprovides encouragement that targeting CDK4/6 may be effica-cious in treatment refractory HER2-amplified disease. These dataalso support the hypothesis that ER and downstreamRb signalingare, at least partially, driving the progression of HER2þ/ERþ

tumors. Clinical data suggest that HER2þ/ERþ and HER2þ/ER�

breast cancers are clinically distinct subgroups based on theirprognosis and response to therapy (37, 38). Furthermore, pre-clinical studies show that crosstalk exists between ER and HER2signaling and activation of either pathway is associated withresistance to targeting the other pathway (39–41). These datasupport the design of the MonarcHER trial (NCT02675231),which will evaluate the combination of abemaciclib plus trastu-zumab with or without hormonal blockade (fulvestrant) inHER2þ/ERþ breast cancer patients.

The use of CDK4/6 inhibitors beyond ERþ/HER2� breastcancer also requires investigation of combinations with stan-dard-of-care cytotoxic chemotherapies. There are preclinical datato suggest that the mechanism of action of molecules that arrestcell cycle may be antagonistic when used with antimitotic agents(35, 42). To address this question, we first investigated the in vivoactivity of abemaciclib in combination with chemotherapy (doc-etaxel) inHER2-amplfied and TNBC xenografts. In TNBC, cotreat-ment with abemaciclib and docetaxel did not antagonize antitu-mor activity of either single agent. Moreover, the addition ofdocetaxel to a triple (anti-HER2, anti-ER, and anti-CDK4/6)combination in HER2þ/ERþ xenografts increased the degree oftumor regressions observed. Analysis of tumor tissues confirmedthat docetaxel did not block the induction of the cell-cycle arrestby abemaciclib. In vitro drug-sequencing experiments confirmedthat induction of growth arrest by abemaciclib does not block theinhibition of cell proliferation or the apoptosis induced bydocetaxel or carboplatin. In fact, either simultaneous or sequen-tial combination of abemaciclib plus either cytotoxic leads to

greater inhibition of tumor cell growth compared with single-agent treatment. The combined activity observed between thesetwo classes of molecules could possibly be explained by fact thatnot all tumor cells will be uniformly arrested by abemaciclibleaving subpopulations of cells sensitive to the mechanism ofaction of antimitotics. It has previously been shown that abema-ciclib can be combined with gemcitabine without antagonism inlung cancer xenografts with either sequential or simultaneousadministration (27). Our in vitro experiments also show thatabemaciclib appears to be unique among the CDK4/6 inhibitorstested. In contrast to abemaciclib, both palbociclib and ribociclibpretreatment significantly reduced the induction of apoptosisby docetaxel in a dose-dependent manner. The specific mechan-isms by which abemaciclib differs from palbociclib and ribociclibin this setting requires further investigation; however, thesedata are encouraging for the development of abemaciclib inindications where combination with cytotoxic chemotherapiesmight be required.

Sensitivity to abemaciclib was also detected in a specific sub-group of TNBC cell lines matching the biomarker profile of intactRb signaling, despite expressing low levels of ER andHER2.Of the17 TNBC cell lines tested, the top 5 most sensitive cell lines hadhigh levels of total and phosphorylated levels of Rb accompaniedby low baseline levels of p16. Abemaciclib IC50 values for thissubset of TNBC cell lines are similar to those measured for theERþ/HER2� and HER2þ/ERþ cell lines. Molecular alterationsassociated with resistance to CDK4/6-targeted therapy (43, 44)through activation of Rb signaling independently of CDK4/6,such as high p16 protein levels or amplification of cyclin E, werefound in those TNBC cell lines least sensitive to abemaciclib.Xenograft studies confirmed that TNBC cell lines selected for highbaseline levels of pRb and low p16 were sensitive to abemaciclibwhereas xenografts with lowpRb andhigh p16were unresponsiveto this therapy. Reduction in pRb levels or RB1 via siRNA knock-down significantly reduced the sensitivity of TNBC cells to abe-maciclib. These data support the hypothesis that Rb dependencegoes beyondER-driven breast cancers and that CDK4/6 inhibitionmay be effective in other subtypes with the appropriate cyclinD:CDK4/6:Rb activation profile. It is likely that the strong associ-ationbetweenER status and abemaciclib response inbreast cancermodels and patients (24) is due to the fact that ER is a usefulsurrogate marker of downstream dependence on cyclinD:CDK4/6:Rb:p16 signaling. Selection of patients based on measurementof intact Rb pathway signaling as characterized by high tumorlevels of pRb, low p16, and an absence of cyclin E1 amplification,may identify additional populations of patients that could benefitfrom abemaciclib therapy and could provide therapeutic benefitin a subset of TNBCswhere there are no current, approved targetedtherapies.

Of interest, additional factors that associate with response toabemaciclib in this study include the presence of high levels ofandrogen receptor (AR) and the presence of PIK3CA- activatingmutations. It should be noted, however, that these factors trackwith the luminal breast cancer subtype and as such may not beindependent biomarkers of response (19, 45). High levels ofPTEN protein associated with abemaciclib response indicate thatintact PI3K pathway signaling may increase the likelihood ofresponse to CDK4/6-targeted therapy.

PD markers, for the early detection of response to abemaciclibtherapy, were also identified in this study using a set of cell-cycle–regulated genes. Expression of a core set of five E2F-regulated






genes (RRM2, TOPO2A, MKI67, MCM7, and CDK2), consistentlytracked with response to abemaciclib in a dose-dependent man-ner across three independent xenograft studies. Furthermore, thedynamic changes in expression of these transcripts trackedwith anincrease in antitumor response induced by abemaciclib whencombined with fulvestrant, tamoxifen, or trastuzumab. The ulti-mate utility of this marker set should now be evaluated in tissuesfrom abemaciclib clinical trials.

In this study, we present a comprehensive preclinical profile ofabemaciclib used as a single agent or in combination withstandard-of-care therapy in each of the known therapeutic molec-ular subtypes of breast cancer. These data further support theclinical use of abemaciclib as monotherapy as well as a possiblecombination partner in ERþ/HER2�, HER2þ/ERþ, and someTNBCs. It is likely that beyond ERþ/HER2� breast cancer, mea-surement of an on-mechanism, validated set of biomarkers toconfirm dependence on Rb -signaling will likely help better selectpatients that will benefit from abemaciclib-based therapies.

Disclosure of Potential Conflicts of InterestR. Beckmann has ownership interest (including patents) in Eli Lilly and

Company. S. Hurvitz reports receiving commercial research grant from Lilly.C. Mockbee is an employee of Eli Lilly. D.J. Slamon is a consultant/advisoryboard member for Eli Lilly. No conflicts of interest were disclosed by the otherauthors.

Authors' ContributionsConception and design: N. O'Brien, D. Conklin, R. Beckmann, C. Mockbee,D.J. SlamonDevelopment of methodology: D. Conklin, K. Chau, E. von EuwAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): T. Luo, K. Chau, J. Thomas, A. Mc Nulty, C. Marchal,E. von EuwAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): N. O'Brien, D. Conklin, R. Beckmann, A. Mc Nulty,S. Hurvitz, C. Mockbee, D.J. SlamonWriting, review, and/or revision of the manuscript: N. O'Brien, R. Beckmann,K. Chau, O. Kalous, S. Hurvitz, C. Mockbee, D.J. SlamonAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): T. Luo, K. Chau

AcknowledgmentsThis study was supported by UCLA funding and a DOD Innovator Award

W81XWH-11-1-0104 to Dennis J. Slamon.

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 indicatethis fact.

ReceivedMarch 29, 2017; revised November 16, 2017; accepted February 16,2018; published first February 26, 2018.

References1. Howell A, Osborne CK, Morris C, Wakeling AE. ICI 182,780 (Faslodex):

development of a novel, "pure" antiestrogen. Cancer 2000;89:817–25.2. Osborne CK. Tamoxifen in the treatment of breast cancer. N Engl J Med

1998;339:1609–18.3. Osborne CK. Aromatase inhibitors in relation to other forms of endocrine

therapy for breast cancer. Endocr Relat Cancer 1999;6:271–6.4. Slamon DJ, Leyland-Jones B, Shak S, Fuchs H, Paton V, Bajamonde A, et al.

Use of chemotherapy plus a monoclonal antibody against HER2 formetastatic breast cancer that overexpresses HER2. N Engl J Med 2001;344:783–92.

5. Verma S, Miles D, Gianni L, Krop IE, Welslau M, Baselga J, et al. Trastu-zumab emtansine for HER2-positive advanced breast cancer. N Engl J Med2012;367:1783–91.

6. Cuzick J, Sestak I, BaumM, Buzdar A, Howell A, Dowsett M, et al. Effect ofanastrozole and tamoxifen as adjuvant treatment for early-stage breastcancer: 10-year analysis of the ATAC trial. Lancet Oncol 2010;11:1135–41.

7. Vogel CL, CobleighMA, Tripathy D, Gutheil JC, Harris LN, Fehrenbacher L,et al. Efficacy and safety of trastuzumab as a single agent in first-linetreatment of HER2-overexpressing metastatic breast cancer. J Clin Oncol2002;20:719–26.

8. Berns K, Horlings HM, Hennessy BT, Madiredjo M, Hijmans EM, Beelen K,et al. A functional genetic approach identifies the PI3K pathway as a majordeterminant of trastuzumab resistance in breast cancer. Cancer Cell2007;12:395–402.

9. Goel S, Wang Q, Watt AC, Tolaney SM, Dillon DA, Li W, et al. Overcomingtherapeutic resistance in HER2-positive breast cancers with CDK4/6 inhi-bitors. Cancer Cell 2016;29:255–69.

10. Miller TW, Balko JM, Fox EM, Ghazoui Z, Dunbier A, Anderson H, et al.ERalpha-dependent E2F transcription can mediate resistance to estrogendeprivation in human breast cancer. Cancer Discov 2011;1:338–51.

11. Miller TW, Hennessy BT, Gonzalez-Angulo AM, Fox EM,Mills GB, ChenH,et al. Hyperactivation of phosphatidylinositol-3 kinase promotes escapefrom hormone dependence in estrogen receptor-positive human breastcancer. J Clin Invest 2010;120:2406–13.

12. O'Brien NA, Browne BC, Chow L, Wang Y, Ginther C, Arboleda J, et al.Activated phosphoinositide 3-kinase/AKT signaling confers resistance totrastuzumab but not lapatinib. Mol Cancer Ther 2010;9:1489–502.

13. Cristofanilli M, Turner NC, Bondarenko I, Ro J, Im SA, Masuda N, et al.Fulvestrant plus palbociclib versus fulvestrant plus placebo for treatment of

hormone-receptor-positive, HER2-negative metastatic breast cancer thatprogressed on previous endocrine therapy (PALOMA-3): final analysis ofthemulticentre, double-blind, phase 3 randomised controlled trial. LancetOncol 2016;17:425–39.

14. Finn RS, Martin M, Rugo HS, Jones S, Im SA, Gelmon K, et al. Palbo-ciclib and letrozole in advanced breast cancer. N Engl J Med 2016;375:1925–36.

15. Hortobagyi GN, Stemmer SM, Burris HA, Yap YS, Sonke GS, Paluch-Shimon S, et al. Ribociclib as first-line therapy for HR-positive, advancedbreast cancer. N Engl J Med 2016;375:1738–48.

16. Prall OW, Sarcevic B, Musgrove EA, Watts CK, Sutherland RL. Estrogen-induced activation of Cdk4 and Cdk2 during G1-S phase progression isaccompanied by increased cyclin D1 expression and decreased cyclin-dependent kinase inhibitor association with cyclin E-Cdk2. J Biol Chem1997;272:10882–94.

17. Harbour JW, Luo RX, Dei Santi A, Postigo AA, Dean DC. Cdk phosphor-ylation triggers sequential intramolecular interactions that progressivelyblock Rb functions as cells move through G1. Cell 1999;98:859–69.

18. Timms JF, White SL, O'Hare MJ, Waterfield MD. Effects of ErbB-2 over-expression on mitogenic signalling and cell cycle progression in humanbreast luminal epithelial cells. Oncogene 2002;21:6573–86.

19. Cancer Genome Atlas N. Comprehensive molecular portraits of humanbreast tumours. Nature 2012;490:61–70.

20. Foster SA, Wong DJ, Barrett MT, Galloway DA. Inactivation of p16 inhumanmammary epithelial cells byCpG islandmethylation.MolCell Biol1998;18:1793–801.

21. Gelbert LM, Cai S, Lin X, Sanchez-Martinez C, Del Prado M, Lallena MJ,et al. Preclinical characterization of the CDK4/6 inhibitor LY2835219:in-vivo cell cycle-dependent/independent anti-tumor activities alone/incombination with gemcitabine. Invest New Drugs 2014;32:825–37.

22. Patnaik A, Rosen LS, Tolaney SM, Tolcher AW, Goldman JW, Gandhi L,et al. Efficacy and safety of abemaciclib, an inhibitor of CDK4 and CDK6,for patients with breast cancer, non-small cell lung cancer, and other solidtumors. Cancer Discov 2016;6:740–53.

23. Finn RS, Crown JP, Ettl J, Schmidt M, Bondarenko IM, Lang I, et al. Efficacyand safety of palbociclib in combination with letrozole as first-line treat-ment of ER-positive, HER2-negative, advanced breast cancer: expandedanalyses of subgroups from the randomized pivotal trial PALOMA-1/TRIO-18. Breast Cancer Res 2016;18:67.

O'Brien et al.





24. Finn RS, Dering J, Conklin D, Kalous O, Cohen DJ, Desai AJ, et al. PD0332991, a selective cyclin D kinase 4/6 inhibitor, preferentially inhibitsproliferation of luminal estrogen receptor-positive human breast cancercell lines in vitro. Breast Cancer Res 2009;11:R77.

25. Turner NC, Ro J, Andre F, Loi S, Verma S, Iwata H, et al. Palbociclib inhormone-receptor-positive advanced breast cancer. N Engl J Med 2015;373:209–19.

26. O'Brien NA, McDonald K, Tong L, von Euw E, Kalous O, Conklin D, et al.Targeting PI3K/mTOR overcomes resistance to HER2-targeted therapyindependent of feedback activation of AKT. Clin Cancer Res 2014;20:3507–20.

27. Konecny GE, PegramMD, Venkatesan N, Finn R, Yang G, RahmehM, et al.Activity of the dual kinase inhibitor lapatinib (GW572016) against HER-2-overexpressing and trastuzumab-treated breast cancer cells. Cancer Res2006;66:1630–9.

28. Tate SC, Cai S, Ajamie RT, Burke T, Beckmann RP, Chan EM, et al. Semi-mechanistic pharmacokinetic/pharmacodynamic modeling of the anti-tumor activity of LY2835219, a new cyclin-dependent kinase 4/6inhibitor, in mice bearing human tumor xenografts. Clin Cancer Res2014;20:3763–74.

29. Infante JR, Cassier PA, Gerecitano JF, Witteveen PO, Chugh R, Ribrag V,et al. A phase I study of the cyclin-dependent kinase 4/6 inhibitor ribociclib(LEE011) in patients with advanced solid tumors and lymphomas. ClinCancer Res 2016;22:5696–705.

30. DeMichele A, Clark AS, Tan KS, Heitjan DF, Gramlich K, GallagherM, et al.CDK 4/6 inhibitor palbociclib (PD0332991) in Rbþ advanced breastcancer: phase II activity, safety, and predictive biomarker assessment. ClinCancer Res 2015;21:995–1001.

31. Dickler MN TS, Rugo HS, Cortes J, Di�eras V, Patt DA, Wildiers H, et al.Results from a phase II study of abemaciclib, a CDK4 and CDK6inhibitor, as monotherapy, in patients with HRþ/HER2�breast cancer,after chemotherapy for advanced disease. J Clin Oncol 2016;34(suppl):abstract 510.

32. Sledge GW Jr, ToiM,Neven P, Sohn J, Inoue K, Pivot X, et al. MONARCH2:abemaciclib in combination with fulvestrant in women with HRþ/HER2�advanced breast cancer who had progressed while receiving endo-crine therapy. J Clin Oncol 2017;35:2875–84.

33. Lenferink AEG, Busse D, Flanagan WM, Yakes FM, Arteaga CL. ErbB2/neukinase modulates cellular p27(Kip1) and cyclin D1 through multiplesignaling pathways. Cancer Res 2001;61:6583–91.

34. Choi YJ, Li XY, Hydbring P, Sanda T, Stefano J, Christie AL, et al. Therequirement for cyclin D function in tumor maintenance. Cancer Cell2012;22:438–51.

35. Roberts PJ, Bisi JE, Strum JC, Combest AJ, Darr DB, Usary JE, et al. Multipleroles of cyclin-dependent kinase 4/6 inhibitors in cancer therapy. J NatlCancer I 2012;104:476–87.

36. Witkiewicz AK, Cox D, Knudsen ES. CDK4/6 inhibition provides a potentadjunct to Her2-targeted therapies in preclinical breast cancer models.Genes Cancer 2014;5:261–72.

37. Hurvitz SA, Andre F, Jiang Z, Shao Z, Mano MS, Neciosup SP, et al.Combination of everolimus with trastuzumab plus paclitaxel as first-linetreatment for patients with HER2-positive advanced breast cancer(BOLERO-1): a phase 3, randomised, double-blind, multicentre trial.Lancet Oncol 2015;16:816–29.

38. Nahta R, O'Regan RM. Therapeutic implications of estrogen receptorsignaling in HER2-positive breast cancers. Breast Cancer Res Treat 2012;135:39–48.

39. Gluck S, Arteaga CL, Osborne CK. Optimizing chemotherapy-free survivalfor the ER/HER2-positive metastatic breast cancer patient. Clin Cancer Res2011;17:5559–61.

40. Osborne CK, Shou J, Massarweh S, Schiff R. Crosstalk between estrogenreceptor and growth factor receptor pathways as a cause for endocrinetherapy resistance in breast cancer. Clin Cancer Res 2005;11(2 Pt 2):865s–70s.

41. XiaW, Bacus S, Hegde P, Husain I, Strum J, Liu L, et al. Amodel of acquiredautoresistance to apotent ErbB2 tyrosine kinase inhibitor and a therapeuticstrategy to prevent its onset in breast cancer. Proc Natl Acad Sci U S A 2006;103:7795–800.

42. Dean JL, McClendon AK, Knudsen ES. Modification of the DNA damageresponse by therapeutic CDK4/6 inhibition. J Biol Chem 2012;287:29075–87.

43. Konecny GE, Winterhoff B, Kolarova T, Qi J, Manivong K, Dering J, et al.Expression of p16 and retinoblastoma determines response to CDK4/6inhibition in ovarian cancer. Clin Cancer Res 2011;17:1591–602.

44. Lukas J, Herzinger T, Hansen K, Moroni MC, Resnitzky D, Helin K, et al.Cyclin E-induced S phase without activation of the pRb/E2F pathway.Genes Dev 1997;11:1479–92.

45. Park S, Koo J, Park HS, Kim JH, Choi SY, Lee JH, et al. Expression ofandrogen receptors in primary breast cancer. Ann Oncol 2010;21:488–92.






2018;17:897-907. Published OnlineFirst February 26, 2018.Mol Cancer Ther Neil O'Brien, Dylan Conklin, Richard Beckmann, et al. Antimitotic and Targeted Therapies in Breast CancerPreclinical Activity of Abemaciclib Alone or in Combination with

Updated version

10.1158/1535-7163.MCT-17-0290doi:

Access the most recent version of this article at:

Material

Supplementary

http://mct.aacrjournals.org/content/suppl/2018/02/24/1535-7163.MCT-17-0290.DC1

Access the most recent supplemental material at:

Cited articles

http://mct.aacrjournals.org/content/17/5/897.full#ref-list-1

This article cites 44 articles, 19 of which you can access for free at:

Citing articles

http://mct.aacrjournals.org/content/17/5/897.full#related-urls

This article has been cited by 6 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://mct.aacrjournals.org/content/17/5/897To request permission to re-use all or part of this article, use this link



http://mct.aacrjournals.org/lookup/doi/10.1158/1535-7163.MCT-17-0290

http://mct.aacrjournals.org/content/suppl/2018/02/24/1535-7163.MCT-17-0290.DC1

http://mct.aacrjournals.org/content/17/5/897.full#ref-list-1

http://mct.aacrjournals.org/content/17/5/897.full#related-urls

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

mailto:[email protected]

http://mct.aacrjournals.org/content/17/5/897


Documents

Preclinical Activity of Abemaciclib Alone or in ... · Neil O'Brien1, Dylan Conklin1, Richard Beckmann2,Tong Luo1, Kevin Chau1, Josh Thomas 1 , Ann Mc Nulty 2 , Christophe Marchal