activated d16her2 homodimers and src kinase mediate … · therapeutics, targets, and chemical...

Therapeutics, Targets, and Chemical Biology

Activated d16HER2 Homodimers and SRC Kinase MediateOptimal Efficacy for Trastuzumab

Lorenzo Castagnoli1, Manuela Iezzi2, Gaia C. Ghedini1, Valentina Ciravolo1, Giulia Marzano1,Alessia Lamolinara2, Roberta Zappasodi3, Patrizia Gasparini4, Manuela Campiglio1, Augusto Amici5,Claudia Chiodoni6, Arianna Palladini7, Pier Luigi Lollini7, Tiziana Triulzi1, Sylvie Menard1, Patrizia Nanni7,Elda Tagliabue1, and Serenella M. Pupa1

AbstractA splice isoform of the HER2 receptor that lacks exon 16 (d16HER2) is expressed in many HER2-positive breast

tumors, where it has been linked with resistance to the HER2-targeting antibody trastuzumab, but the impact ofd16HER2 on tumor pathobiology and therapeutic response remains uncertain. Here, we provide genetic evidencein transgenic mice that expression of d16HER2 is sufficient to accelerate mammary tumorigenesis and improvethe response to trastuzumab. A comparative analysis of effector signaling pathways activated by d16HER2 andwild-type HER2 revealed that d16HER2 was optimally functional through a link to SRC activation (pSRC).Clinically, HER2-positive breast cancers from patients who received trastuzumab exhibited a positive correlationin d16HER2 and pSRC abundance, consistent with the mouse genetic results. Moreover, patients expressing highpSRC or an activated "d16HER2metagene" were found to derive the greatest benefit from trastuzumab treatment.Overall, our results establish the d16HER2 signaling axis as a signature for decreased risk of relapse aftertrastuzumab treatment. Cancer Res; 74(21); 6248–59. �2014 AACR.

IntroductionHER2 is a 185-kDa transmembrane receptor that belongs to

the HER family of receptor tyrosine kinases (RTK), includingHER1 (EGFR), HER3, and HER4. Binding of specific ligands tothe extracellular domain (ECD) of HER1, HER3, and HER4induces the formation of homo- and heterodimers, with acti-vated HER2 as a preferred partner (1). Overexpression oramplification of ERBB2 (HER2) occurs in 15% to 20% of invasivebreast cancers and is associated with more aggressive disease

and, until the advent of HER2-targeted agents, a worse outcome(2). In the metastatic setting, the addition of HER2-targetedagents to chemotherapy improved disease-free survival (�37%),overall survival (�22%), and overall response rate (�67%; ref. 3).Trastuzumab, a recombinant humanized anti-HER2 monoclo-nal antibody, combined with chemotherapy is a foundation ofcare for patients with HER2-positive breast cancers (2, 3).However, most HER2-positive breast cancer patients who ini-tially respond to trastuzumab subsequently become refractoryand disease progresses. Several intrinsic mechanisms wherebytumors escape HER2 inhibition by trastuzumab have beensuggested (4), including altered forms of HER2 itself (5, 6) andactivating HER2 mutations identified in HER2 gene amplifica-tion–negative breast cancer (7). We and others (8, 9) reportedthat the splice variant of human HER2 lacking exon 16, herenamed d16HER2, and characterized by an imbalance in thenumber of cysteines in the ECD portion and by the constitutivegeneration of stable HER2 homodimers, is a highly penetrantHER2 oncogenic alteration. d16HER2, identified in most humanHER2-positive primary breast cancers, effects a decrease intrastuzumab binding in vitro (9) and promotes resistance totrastuzumab in multiple cell lines (10).

Although transgenic (tg) mousemodels of the rodent form ofHER2 (Neu) have been instrumental in the study of basiconcogene activity (11–14), the inherent limitations of the rodentNeu tg models have led to the development of tg mouse modelsfor the humanwild-type, full-lengthHER2 (WTHER2; refs. 15–17)to study the mechanisms regulating HER2-driven cancer recur-rence, trastuzumab sensitivity, and resistance. However, bothrodent and human HER2 transgenes require activating muta-tions to become oncogenic, implying that genetic changes in

1Molecular Targeting Unit, Department of Experimental Oncology andMolecular Medicine, Fondazione IRCCS, Istituto Nazionale dei Tumori,Milan, Italy. 2AgingResearchCentre, G.D'Annunzio University, Chieti, Italy.3C. Gandini Medical Oncology, BoneMarrow Transplantation Unit, Depart-ment of Experimental Oncology and Molecular Medicine, FondazioneIRCCS, Istituto Nazionale dei Tumori, Milan, Italy. 4Tumor Genomics Unit,Department of Experimental Oncology and Molecular Medicine, Fonda-zione IRCCS, Istituto Nazionale dei Tumori, Milan, Italy. 5Department ofBioscience and Biotechnology, University of Camerino, Camerino, Italy.6Molecular Immunology Unit, Department of Experimental Oncology andMolecular Medicine, Fondazione IRCCS, Istituto Nazionale dei Tumori,Milan, Italy. 7Department of Experimental, Diagnostic and Specialty Med-icine (DIMES), University of Bologna, Bologna, Italy.

Note: Supplementary data for this article are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/).

L. Castagnoli, M. Iezzi, E. Tagliabue, and S.M. Pupa contributed equally tothis article.

Corresponding Author: Serenella M. Pupa, Molecular Targeting Unit,Department of Experimental Oncology and Molecular Medicine, Amadeo-Lab, Fondazione IRCCS Istituto Nazionale dei Tumori, Via Amadeo 42,20133 Milan, Italy. Phone: 39-02-2390-2573; Fax: 39-02-2390-2692;E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-14-0983

�2014 American Association for Cancer Research.

CancerResearch

Cancer Res; 74(21) November 1, 20146248

on March 3, 2020. © 2014 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst August 27, 2014; DOI: 10.1158/0008-5472.CAN-14-0983

http://cancerres.aacrjournals.org/

addition to HER2 overexpression are required for mammarytumorigenesis (17, 18). In that context, we generated a FVBmouse line that transgenically expresses the human d16HER2isoformand stochastically developsmetastaticmultifocalmam-mary tumors expressing heterogeneous levels of constitutivelyactivated stable HER2 homodimers (pd16HER2D); these homo-dimers couple to multiple oncogenic downstream signal trans-duction pathways, including SRC kinase (19).The oncogenic activity and trastuzumab susceptibility of

d16HER2-positive mammary tumors (9, 10, 20), as well as therelationship of d16HER2 with WTHER2-driven pathobiologicand clinical features in human HER2-overexpressing breastcancers, await clarification.Here, we provide evidence in both mice and humans that

d16HER2-positive tumors respond significantly to trastuzu-mab and that this response depends on the functional rela-tionship and coexpression of activated d16HER2 stable homo-dimers and SRC kinase.

Materials and MethodsTumor cell linesThe d16HER2- and WTHER2-positive mammary tumor cell

lines MI6 and WTHER2 were established from spontaneousprimary mammary carcinomas of an 18-week-old virgin FVB-d16HER2 and a 34-week-old virgin FVB-huHER2 tg femalemouse, respectively. Briefly, primarymammary tumors excisedfrom sacrificed mice were finely minced, incubated in eryth-rocyte lysis buffer, enzymatically digested (Collagenase/Hyal-uronidase; STEMCELL Technologies) and extensively washedbefore examination in four high-power fields based on trypanblue staining (see Supplementary Fig. S1A and S1B for cellmembrane expression of d16HER2 andWTHER2, respectively).In the case of d16HER2-positive tumors, whole mammarytumor cell suspensions were selectively separated under sterileconditions by AutoMACS separator (Miltenyi Biotec) to obtainhomogenous EPCAM- and d16HER2-positive neoplastic cellcultures (manuscript in preparation). The MI6 and WTHER2cell lines were maintained in complete culture medium (Mam-moCult; STEMCELL Technologies) supplemented with 1%fetal bovine serum (FBS; Sigma) and penicillin–streptomycin(Sigma-Aldrich) and cultured at 37�C in a 5% CO2 atmosphere.MI6 and WTHER2 tumor cell lines were routinely tested byflow cytometry and quantitative real-time PCR (qRT-PCR).

Tg mice and in vivo therapyA breeding colony of FVB d16HER2 tg mice (FVB/NHsd-

Tg(D16HER2-LUC)6157Acam) was generated as described(19) and bred in the Animal Facility of Fondazione IRCCSIstituto Nazionale dei Tumori. Animal care and experimentalprocedureswere approved by the Ethics Committee for AnimalExperimentation of the Institute according to the Italian law.DNA extracted from tail biopsies was used for routine geno-typing by PCR analysis (primers: F, 50-GGCTCAGTGACC-TGTTTTGG-30 and R, 50-TGATGAGGATCCCAAAGACC-30),with an expected amplicon length of 231 bp.Micewere inspect-ed twice weekly by palpation.FVB-huHER2 (WTHER2) transgenic mouse line (FVB/N-Tg

(MMTV-ERBB2)5Erick) (Fo5) carries the full-length normal

huHER2 gene under the control of the MMTV promoter onan FVB background (17) and was obtained from Genentech,Inc. FVB-huHER2 mice were bred in animal facilities of theDIMES Department of the University of Bologna and geneti-cally screened by PCR using a primer set specific to humangrowth hormone exons 4 and 5 included in the transgenebackbone (17). Mice were inspected weekly by palpation. Invivo experiments were performed in compliance with theItalian and European guidelines and were approved by theInstitutional Review Board of the University of Bologna. Pro-gressively growing masses�50 mm3 were scored as tumors inboth tg models. Susceptibility of d16HER2 to trastuzumabtreatments was assessed in d16HER2-positive tg spontaneousand in orthotopic d16HER2 and WTHER2-positive models. Inthe first set of in vivo experiments, d16HER2 tg mice wereinjected i.p. with trastuzumab (Roche) or diluent NaCl solution(0.9%) in a short (n¼ 8/group) and prolonged (n¼ 7–8/group)administration protocol. In the short treatment, tg mice weretreated with trastuzumab (8 mg/kg) once per week for 5 weeksstarting from 8 weeks, when only microscopic tumor lesionswere present (19), until 12 weeks of age. The study wasterminated at 29 weeks of age, when all d16HER2 mice devel-oped the first spontaneous tumor. In the prolonged protocol,d16HER2 tgmice received trastuzumab (4mg/kg) twiceweeklyfrom 8 until 42 weeks of age. In each experiment, tumors werecalibrated twice weekly and tumor volume was calculated as0.5 � d1

2 � d2, where d1 and d2 are the smaller and largerdiameters, respectively. FVB female mice (6–8-weeks old; bodyweight, 20–25 g) were purchased from Charles River Labora-tories. Mice (n ¼ 10/group) were injected into the mammaryfat pad (m.f.p.) with 1� 106MI6 orWTHER2 tumor cells.Whentumors reached 50 mm3, mice were randomized into twogroups to receive biweekly i.p. injections of 4 mg/kg trastuzu-mab or diluent NaCl solution (0.9%). The use of the twodifferent dosing schedules of trastuzumab administration isbased on the reliable results we previously obtained (21, 22).Tumors were calibrated twice weekly and tumor volume wascalculated as above. Mice were sacrificed when tumor volumesreached �2,000 mm3. Each tumor specimen was placed intoliquid nitrogen for biochemical analyses. For histopathologicanalyses, tumors and lungs were fixed overnight in 10% neu-tral-buffered formalin and transferred into 70% ethanol beforeprocessing and paraffin embedding. Paraffin sections (5-mmthick) were stained with hematoxylin and eosin (H&E). Lungmetastases were induced with 105 and 106 viable MI6 andWTHER2 tumor cells, respectively, injected i.v. in 0.4mL of PBSin FVB female mice. Mice were randomized into two groups (n¼ 8/group) to receive biweekly i.p. injections of 4 mg/kgtrastuzumab or diluent NaCl solution 0.9%, respectively. Treat-ment started 7 days after cell injection. Mice were sacrificedand necropsied 11 weeks after d16HER2 and WTHER2 cellinjection. Lungs were perfused with black India ink to outlinemetastases and fixed in Fekete's solution. Lung metastaseswere counted using a dissection microscope.

qRT-PCROf 84 HER2-positive human breast cancer specimens, 43

frozen primary breast cancer were available for analysis by

Activated d16HER2/SRC Axis Predicts Trastuzumab Benefit

www.aacrjournals.org Cancer Res; 74(21) November 1, 2014 6249




qRT-PCR to determine the amount of d16HER2 transcript, asnormalized to the amount of WTHER2 mRNA. Total RNAfrom human primary breast cancer frozen specimens wasextracted with TRIzol (Invitrogen) according to the manu-facturer's instructions. cDNAs were reverse-transcribedfrom 1 mg of total RNA in a 20-mL volume with SuperScriptIII (Invitrogen) using random-hexamer primers and exam-ined by qRT-PCR using Applied Biosystems SYBR Greendye–based PCR assay on the ABI Prism 7900HT sequencedetection system (Applied Biosystems). d16HER2 andWTHER2 isoforms were amplified using 200 nmol/L primers(10). Data were normalized toGAPDH (23). Relative abundanceof d16HER2 mRNA compared with that of WTHER2 wascalculated by the comparative Ct method (24), with d16HER2transcript levels indicated as the ratio 2�(DCt)d16HER2/2�(DCt)WTHER2. To correlate d16HER2 transcript and pSRCexpression levels in human breast cancers, gene expressiondatawere split in twogroups according to tertiles: low, contain-ing values under the first tertile, and high, containing valuesgreater than the first tertile.

Statistical analysesDifferences in tumor multiplicity curves in both d16HER2

and WTHER2 tg models and differences in trastuzumab anti-tumor activity in orthotopic MI6 and WTHER2 models werecalculated, by the two-tailed unpaired t test. Differences wereconsidered significant at P < 0.05.

Linear regression and Pearson correlation coefficient rwerecalculated to estimate the correlation of (i) pd16HER2M andpd16HER2D with pSRC and of pWTHER2 with pSRC levelsboth under nonreducing and reducing conditions in proteinextracts from both d16HER2 and WTHER2 tg models; (ii)d16HER2 with WTHER2 gene expression levels; and (iii) pSRC(%) with d16HER2 transcript levels in human primary breastcancers.

Survival was assessed using the Kaplan–Meier estimator,whereas log-rank test was used to compare survival distribu-tions. Survival analysis was carried out using Cox proportionalhazards regression models, and the effects of explanatoryvariables on event hazard were quantified by hazard ratios(HR; ref. 25).

Data for the "activated-d16HER2 metagene," constructedbased on the Illumina Whole-Genome DASL gene expressionprofiling of 21 HER2-positive breast cancers characterizedfor pSRC and d16HER2 expression (GSE55348, see Supple-mentary Table S4), were quantile-normalized using Bead-Studio software and filtered with a data matrix containing22,121 probes, corresponding to 15,715 Entrez Ids. Pathwaysdifferentially enriched in activated d16HER2 tumors wereevaluated by Gene Set Enrichment Analysis using GSEAv2.0.13 (26) on a 193-cancer-related gene set (24). Permuta-tion type was applied 1,000 times. Core members of eachsignificantly (P < 0.05) enriched gene set were extracted andtheir mean expression levels were considered as the "acti-vated-d16HER2 metagene" value. The metagene was calcu-lated in HER2-positive breast cancer biopsies of two publiclyavailable datasets, GSE22358 (27) and GSE41656 (28), forwhich pathologic complete response information was avail-

able. Differences in "activated-d16HER2 metagene" valuesbetween responders and nonresponders were evaluated bythe unpaired t test. Area under the ROC curve was calculatedby nonparametric ROC analysis (29).

ResultsPathobiologic characteristics of mouse linestransgenically expressing human d16HER2 andWTHER2

We first investigated the oncogenicity driven by d16HER2and WTHER2 in tg models. Kaplan–Meier disease-free sur-vival analysis (Fig. 1A) clearly revealed the significant sur-vival advantage (P < 0.001) of WTHER2 compared with thed16HER2 variant. Indeed, mammary tumors in tg WTHER2virgin females (n ¼ 40) arose after 8 months of age andprogressively thereafter only in 85% of mice, whereas alld16HER2 tg virgin females (n ¼ 87) developed multipleasynchronous mammary tumors between 8 and 32 weeksof age. Tumor multiplicity (Fig. 1B) was also significantlyhigher in d16HER2 tg mice (P < 0.001), with a mean numberof five lesions at 30 weeks of age (n ¼ 45) versus a mean oftwo in WTHER2 females at 60 weeks (n ¼ 39).

FISH analysis to evaluate the genetic status of HER2 in exvivo d16HER2 and WTHER2-positive tumor cells derived fromthe spontaneous tg corresponding lesions (Fig. 1C) revealed asingle FISH signal on two chromosomes, both in metaphasespreads and in interphase nuclei (arrows) from d16HER2-positive tumor cells (left), whereas in WTHER2-positive cells(right), amplified signals were identified within 2–3 chromo-somes (arrows). Cytogenetic analysis revealed a near-tetra-ploid karyotype (76–88 chromosomes) of WTHER2-overex-pressing cells compared with a diploid karyotype observed ind16HER2-positive cells. Because our tg models are heterozy-gous for d16HER2 andWTHER2, these results suggest selectiveduplication of the chromosome carrying the human transgenein mammary tumor cells derived from tg mice.

Histopathologic analysis of tumors showed that both strainsdevelop mammary ductal adenocarcinomas; however, where-as all d16HER2 tumors (Fig. 1D) and the vast majority ofWTHER2 tumors (Fig. 1E) grew with a solid pattern, someWTHER2 tumors showed papillary differentiation (data notshown). Moreover, whereas WTHER2 tumors were composedof uniform cells growing with a homogeneous solid appearance(Fig. 1E), different zones were detected in d16HER2 tumors: anouter zone composed of cells with an epithelial appearance andpale cytoplasm (�); an intermediate zone formed of fusiformcells with darker nuclei (��); and an inner zone of cells with anepithelial appearance and pinkish cytoplasm (��; Fig. 1D).Immunohistochemical analysis classified both tumors in thesame intrinsic subtype, i.e., ERBB2-overexpressing (30), as theyare E-CADHERIN positive, confirming their ductal type, andexpress only low levels of estrogen receptor (ER), undetectablelevels of progesterone receptor (PR), high levels of the prolifer-ation marker PCNA (>14%), and a positivity for HER2 on mosttumor cells (Fig. 1D and E). Interestingly, the intensity anddistribution ofHER2 expression differed considerably in the twostrains. WTHER2 tumors showed strong and uniform mem-brane staining for HER2-tg protein on most tumor cells, with

Castagnoli et al.

Cancer Res; 74(21) November 1, 2014 Cancer Research6250




only a slight increase at the edges of the tumors (Fig. 1E),whereas in d16HER2 tumors, membrane staining for d16HER2tg protein was especially strong on the outer zone (�), fadedin the intermediate zone (��), and again well detectable in theinner zone (��; Fig. 1D).

Trastuzumab-mediated antitumor activity in d16HER2and WTHER2 preclinical modelsTo address the critical controversy about trastuzumab

susceptibility, we performed a series of in vivo therapeuticbioassays using trastuzumab in d16HER2 tg mice and in FVBmice orthotopically transplanted with MI6 d16HER2-posi-tive and WTHER2-positive tumor cell lines (Fig. 2). Trastu-zumab treatment of tg mice for either a short (Fig. 2A and B)or prolonged (Fig. 2C and D) time starting at 8 weeks of age,when only microscopic tumor lesions are present (19), led inboth cases to a significant reduction in mammary tumorincidence (P ¼ 0.0038, short; P ¼ 0.0065, prolonged) andtumor multiplicity (��, P ¼ 0.0004, short; ��, P ¼ 0.0002,prolonged) as compared with the control groups, suggesting

a clear survival advantage upon trastuzumab treatment. Inthe prolonged trastuzumab treatment, 1 out of 7 treated tgmice was completely protected until the 42nd week ofobservation (Fig. 2C). The validity of d16HER2- (MI6) andWTHER2-positive cancer cells grown in the m.f.p. of parentalFVB females as appropriate therapeutic models, especiallyuseful for WTHER2-positive tumors that typically have along latency, was confirmed by histologic examination andHER2 staining of MI6 and WTHER2 orthotopic transplantsand their spontaneous tg primary tumor of origin; both thed16HER2 and WTHER2 m.f.p. models (Fig. 2E, left) strictlyrecapitulated histologic and immunohistochemical featuresof spontaneous primary mammary tumors (Fig. 2E, right),reproducing both the morphologic differentiation and dif-ferences in HER2 expression. Moreover, flow cytometry toassess expression levels of d16HER2 and WTHER2 forms inthe corresponding tumor cell lines showed a lower MFI ofd16HER2-positive cells than that of WTHER2-positive cells(Supplementary Fig. S1), consistent with the HER2 stainingpattern (Fig. 1E).

Figure 1. Pathobiologiccharacteristics of mouse linestransgenically expressing humand16HER2 and WTHER2 receptors.A, tumor-free survival of d16HER2and WTHER2 tg mice. P value bylog-rank test. B, mean number ofpalpable mammary carcinomasdeveloped in d16HER2 (!) andWTHER2 (^) tg lines. Data aremean � SEM. ��, P < 0.001 byunpaired t test. C, FISH analysis ofa metaphase spread and aninterphase nucleus obtained fromex vivo d16HER2 (left) and on ametaphase spread of WT (right)HER2-positive mammary tumorcells derived from thecorresponding spontaneoustumor. D and E, histologic andimmunohistochemical analysis ofprimary tumors from d16HER2 andWTHER2 tg mice, respectively.H&E staining showed three zonesin d16HER2 tumors: an outer zone(pale cells, �), an intermediate zone(darker fusiform cells, ��), and aninner zone (pinkish cytoplasm,��), as compared with ahomogeneously solid and uniformappearance of WTHER2 tumors.Magnification, �200. Consistentwith the histologic appearance,E-CADHERIN (E-CAD) and HER2positivity were more marked in theouter and inner zones in d16HER2tumors and uniformly and stronglypositive in WTHER2 tumors.Proliferative activity, indicated byPCNA positivity, was similar in thetwo tumors. Estrogen (ER) andprogesterone receptor (PR)staining was negative in bothtumors.






We then tested the therapeutic activity of trastuzumabin parental FVB females (n ¼ 10/group) orthotopicallyimplanted with MI6 (Fig. 2F) and WTHER2 (Fig. 2G) cells.Trastuzumab treatment was started when mammary tumorsbecame palpable (�50 mm3) and continued until tumor vol-ume reached 2,000 mm3. As compared with controls (n ¼ 10/group), trastuzumab effectively suppressed d16HER2-driventumor growth (P < 0.001; Fig. 2F), whereas the benefits oftrastuzumab in mice with WTHER2 tumors were evident butnot statistically significant (Fig. 2G). Assessment of the effectof trastuzumab treatment on metastases apart from that onthe primary tumors using mice injected i.v. with 105 d16HER2or 106 WTHER2 tumor cells showed that trastuzumab signi-ficantly reduced the metastatic ability of MI6 (89% inhibition)and, to a lesser extent, that of WTHER2 tumor cells

(75% inhibition; Table 1). Together, these results indicate thattrastuzumab can inhibit the oncogenic properties of d16HER2-expressing mammary tumor cells.

Signal transduction axes downstream of d16HER2 andWTHER2 isoforms

Activation of the intrinsic tyrosine kinase activity ofd16HER2 was analyzed by Western blotting both under non-reducing and reducing conditions in eight primary tumorprotein extracts (Fig. 3A and B). The signaling activity down-stream of WTHER2 was analyzed only under reducing condi-tions (n ¼ 9; Fig. 3E), as HER2 stable homodimers were neverdetected in theWTHER2model. Analysis consistently revealedbasal d16HER2 homodimers (d16HER2D) migrating above225 kDa, whose phosphorylation levels (pd16HER2D) were

Figure 2. Trastuzumab-mediatedantitumor activity in d16HER2preclinical models. A, tumor-freesurvival and (B) tumormultiplicity oftg d16HER2 mice treated with ashort trastuzumab protocol (~, 8mg/kg i.p. once weekly for 5 weeks)and diluent saline solution (*, 150mL, i.p. once weekly for 5 weeks).C, tumor-free survival and (D) tumormultiplicity of tg d16HER2 micetreated with a prolongedtrastuzumab protocol (~, 4 mg/kgi.p. twice weekly until sacrifice) anddiluent saline solution (*, 150 mL,i.p. twice weekly until sacrifice).Short and prolonged treatmentsstarted when mice were 8 weeks ofage. Data are mean � SEM.Differences were assessed by log-rank test (A and C) and by unpairedt test (B, ��, P ¼ 0.0004; and D,��, P ¼ 0.0002). E, H&E and HER2staining of MI6 (top) and WT(bottom) HER2 tumor cells injectedin the m.f.p. of parental FVBfemales (left) and of theirspontaneous tg primary mammarytumor of origin (right). F and G,trastuzumab-mediated antitumoractivity in parental FVB micefollowing orthotopic injection ofMI6 and WT HER2-positive tumorcells, respectively. Tumor-bearingmice were treated i.p. withtrastuzumab (~, 4 mg/kg twiceweekly until sacrifice) and diluentsaline solution (*, 150 mL, twiceweekly until sacrifice) in thepresence of evident disease. Dataare mean � SEM. ��, P < 0.001 byunpaired t test.

Castagnoli et al.





particularly marked in four samples (lanes 3, 6, 7, and 8), lessintense in three (lanes 1, 4, and 5), and absent in one (lane 2) ofall tested tumor lysates (Fig. 3A, top and bottom). Constitutivebasal d16HER2D expression was less abundant and moredifficult to resolve than its d16HER2 monomeric counterpart(d16HER2M), while both d16HER2M and D were always sig-nificantly activated within the same tumor sample (Fig. 3A,lanes 3, 6, 7, and 8). This scenario confirms our previousfindings (19) and demonstrates that stable d16HER2D isconstitutively and heterogeneously activated in d16HER2-pos-itive lesions. Analysis of cell signaling downstream of d16HER2and WTHER2 receptors, evaluated under reducing conditions(Fig. 3B and E), revealed that phosphotyrosines of pd16HER2Mand activated WTHER2 (pWTHER2) act as docking sites forproteins, initiating signals that are transduced to the nucleusthrough different circuitries, including the mitogen-activatedprotein kinases (MAPK), AKT, SRC, and STAT3. However, theactivation levels of d16HER2 andWTHER2 signal transductionpathways differed, with higher levels of pd16HER2D andpd16HER2M always significantly coupled to elevated pSRClevels (Fig. 3A–D). This finding strongly suggests the existenceof a pd16HER2D–pSRC signaling axis that amplifies d16HER2-driven oncogenic signals, consistent with a significant directcorrelation between pd16HER2D and pSRC (r ¼ 0.8787; P ¼0.0041; Fig. 3C) and between pd16HER2M and pSRC (r ¼0.8199; P ¼ 0.0127; Fig. 3D). Note that despite very high-levelexpression of native SRC kinase only in WTHER2-positivetumors in all samples, SRC was activated in only 6 out of 9cases independent of pWTHER2 status (Fig. 3E), such that nodirect significant correlation between pWTHER2 and pSRCwas apparent in the WT model (Fig. 3F).To test whether the signal in the d16HER2 model is orches-

trated mainly by downstream pSRC, we examined MI6 cellstreated or not at different times with anti-HER2 ECD MAbsMGR2 and 4D5 and solubilized under nonreducing and reduc-ing conditions; Western blotting showed that activation levelsof d16HER2D, d16HER2M, and SRC in treated cells decreasedin parallel in the same time frame compared with untreatedcells, whereas the phosphorylation status of MAPK and AKTwas kinetically less dependent on the downmodulation ofpd16HER2D, pd16HER2M, and pSRC levels (Fig. 3G and H).Consistent with our biochemical analyses indicating a signif-icant functional direct interaction only between d16HER2 andpSRC (Fig. 3) and with in vitro data demonstrating a physical

interaction between d16HER2 and pSRC (10), the tumor cellmembrane in both d16HER2-positive primary lesions and lungmetastases coexpressed d16HER2 and pSRC, as indicated byIHC, immunofluorescence, and confocal microscopy analyses(Fig. 4).

Correlation of d16HER2 and SRC activity withtrastuzumab-mediated clinical efficacy

To evaluate the potential association of d16HER2 and pSRCin the human setting and to test whether patient outcome aftertrastuzumab treatment might be influenced by d16HER2signaling through pSRC activity, we examined a retrospectiveseries of 84 primary human HER2-positive cases treated adju-vantly with trastuzumab (see Supplementary Table S1 forbreast cancer patient pathobiologic and clinical characteris-tics). Evaluation of pSRC expression in formalin-fixed, paraffin-embedded breast cancer sections by confocal microscopy (Fig.5A and B) revealed high pSRC positivity (>20%) in 34 of 84tumors (Fig. 5B), whereas the remaining 50 breast cancersexpressed pSRC levels ranging from 0% to <20% (Fig. 5A, leftand right, respectively). qRT-PCR analysis of 43 of the 84available frozen breast cancer samples for d16HER2 transcriptlevels, scored as low or high, revealed a significant associationbetween pSRC (%) and d16HER2 (P ¼ 0.0482) expression (Fig.5C). Moreover, tumors with pSRC (%) >0 showed a significantdirect correlation between d16HER2 transcript and pSRCexpression (r¼ 0.6880; P¼ 0.0016; Fig. 5D), strongly suggestingthat the presence of active d16HER2D in primary HER2-positive breast cancers is reflected by high SRC activation.Finally, breast cancer patients with tumors expressing highd16HER2 and pSRC levels exhibited a lower relapse rate aftertrastuzumab treatment than did d16HER2-high/pSRC-low ord16HER2-/pSRC-low patient subgroups (1/12 vs. 9/31). In lightof these results, we revisited the entire 84 case series, in which34 showed high pSRC positivity (>20%); whereas no differencesin clinical-pathobiologic parameterswere found between high-and low-pSRC–expressing tumors (Supplementary Table S2),the relapse-free survival of patients with a high pSRC score intheir primary tumors showed a significantly lower progressivedisease rate after trastuzumab treatment than those with alow pSRC score [HR, 0.28; 95% confidence interval (CI), 0.09–0.83; P ¼ 0.022; Fig. 5E), suggesting that high pSRC levels inearly tumors predicts benefit from trastuzumab-containingtreatment.

Table 1. Effect of trastuzumab treatment on experimental lung metastasis after i.v. injection of MI6 andWTHER2-positive tumor cells

Tumor cell lines

MI6 WTHER2

Treatment Incidence Median Range % Inhibition Incidence Median Range % Inhibition

Control 8/8 211 104–257 8/8 64 4–223Trastuzumab 7/8 24a 0–49 89 7/8 16b 0–43 75

aP < 0.001 by Wilcoxon–Mann–Whitney test in trastuzumab vs. control mouse groups.bP < 0.05 by Wilcoxon–Mann–Whitney test in trastuzumab vs. control mouse groups.






Figure 3. Western blotting analysesof the signal transduction axisdownstream of d16HER2 andWTHER2 forms. A, protein extractsfrom d16HER2 specimens (n ¼ 8)were separated by 3% to 8%gradient SDS-PAGE undernonreducing conditions andprobed with anti-HER2(d16HER2M and d16HER2D, top)and anti-phosphoHER2(pd16HER2M and pd16HER2D,bottom) antibodies. B, the sameprotein extracts were separated by4% to 12% gradient SDS-PAGEunder reducing conditions toevaluate the basal and activationstatus (p) of d16HER2M, SRC,STAT3, AKT, andMAPK. Actin wasused to normalize protein loading.C, linear regression analysis ofpd16HER2D versus pSRC in thed16HER2 protein extracts (seeMaterials andMethods) analyzed inA and B. D, linear regressionanalysis of pd16HER2M versuspSRC in the d16HER2 proteinextracts (see Materials andMethods) analyzed in B. E, proteinextracts from WTHER2 specimens(n ¼ 9) were separated by 4% to12% gradient SDS-PAGE underreducing conditions to evaluate thebasal and activation status (p) ofWTHER2, SRC, STAT3, AKT, andMAPK. Vinculin was used tonormalize protein loading.Autoradiographs of B and E wereacquired at different exposuretimes to obtain optimal imageresolution. F, linear regressionanalysis of pWTHER2 versuspSRC in WTHER2 protein extracts(see Materials and Methods)analyzed in E. G, MI6 proteinextracts from cells treated with theanti-HER2/ECD mAbs MGR2 and4D5 for different times (5 minutes,30 minutes, 4 hours, and 24 hours)were separated by 3% to 8%gradient SDS-PAGE undernonreducing conditions andprobed with anti-HER2(d16HER2M and d16HER2D) andanti-phosphoHER2 (pd16HER2Mand pd16HER2D) antibodies. H,the same protein extracts as inG were separated by 4% to 12%gradient SDS-PAGE underreducing conditions to evaluate thebasal and activation status (p) ofd16HER2M, SRC, AKT, andMAPK. Actin was used tonormalize protein loading.

Castagnoli et al.





To further investigate whether patients with high d16HER2transcript/signaling are those more sensitive to trastuzumab-mediatedHER2 blocking, we generated an "activated-d16HER2signature" by comparing the gene expression profiles of 21 ofthe 43 qRT-PCR–tested breast cancer cases according tod16HER2 and pSRC expression. Tumors expressing d16HER2and pSRC-high were significantly enriched in hypoxia, tumormetastasis, and cell motility pathways in the GSEA analysis(Supplementary Fig. S2).Moreover, ametagene consisting of 73leading genes (Supplementary Table S3) in the enrichment ofthese pathways discriminated, with good performance, caseswith active d16HER2 ("activated-d16HER2 metagene"; AUC ¼0.94; 95% CI, 0.83–1.04; P¼ 0.0039; Fig. 6A). In silico analysis of"activated-d16HER2 metagene" expression in two datasets,GSE22358 (27) and GSE41656 (28), of HER2-positive breastcancer patients treated or not with trastuzumab-based neoad-juvant therapy showed significantly higher expression of thismetagene (P¼ 0.0305) in patients who achieved a complete ornear-complete response to trastuzumab than inpartial respon-ders (Fig. 6B), with a good performance prediction (Fig. 6C). Incontrast, responders and nonresponders to neoadjuvant ther-apy consisting of chemotherapy alone revealed no difference inthe "activated-d16HER2 metagene" expression level (Fig. 6D),strongly suggesting that human breast cancers with high

d16HER2 signaling benefit significantly from the addition oftrastuzumab to chemotherapy treatment.

DiscussionIn this study, we provide evidence that d16HER2 variant

constitutes a more aggressive HER2 isoform susceptible totrastuzumab treatment. The significantly shorter latency andthe consistently higher tumor multiplicity in the d16HER2 tgline as compared with the WTHER2 tg line imply that geneticchanges in addition to WT gene amplification are required formammary tumorigenesis. Moreover, cytogenetic and FISHanalyses of ex vivo d16HER2 tg tumor cells revealed a diploidkaryotype and a single signal in two chromosomes, whereas exvivoWTHER2 tg cancer cells showed amarked aneuploidy andamplified signals on 2–3 chromosomes, supporting the notionof a "firestorm" genomic pattern (31) needed to driveWTHER2-associated mammary tumorigenesis.

About 90% of women with HER2-positive breast cancer andlocally disseminated disease have been reported to coexpressthe oncogenic d16HER2 isoform (9, 10). The coexistence ofd16HER2 with the other two naturally occurring HER2 splicevariants, herstatin and p100, with contrasting roles in tumorcell biology (32), with truncated HER2 isoforms (33) and with

Figure 4. Immunohistochemicaland immunofluorescence analysesof pSRC and HER2 expression inprimary tumor and lung metastasisfrom a tg d16HER2 mouse.Immunohistochemistry showedpSRC and HER2 expression in thesame zones. Confocal microscopyrevealed colocalization of the twoproteins on mammary tumor cellmembranes.






HER2 somatic mutations (7) greatly contributes in complicat-ing the HER2-derived proteome and increasing the heteroge-neity of HER2-positive disease. In this context, it is importantto note that if all the described HER2 forms are driver events,then HER2-positive breast cancer patients might benefit clin-ically from existing HER2-targeted drugs, although this seemsunlikely (34).

It has remained unclear whether d16HER2 is sensitive totrastuzumab treatment (9, 10, 20, 35) and whether d16HER2represents a mechanism of resistance to trastuzumab inpatients with HER2-overexpressing breast cancer (36, 37). Inkeeping with a pilot study of immunodeficient mice injected inthemammary glandwith a d16HER2-positive transfectant (20),we found that spontaneous tumor development in d16HER2 tgmice was significantly impaired by trastuzumab administeredas monotherapy and that prolonged treatment was evencurative in one d16HER2-positive female observed until 42weeks of age. Moreover, tumors formed after m.f.p. injectionsof MI6 and WTHER2 cells showed marked benefit of trastu-zumab treatment only in d16HER2-positive tumors, whereasWT tumors benefited onlymoderately. Also, the antimetastaticeffects of trastuzumab on experimental lung metastases

induced by MI6 and WTHER2-positive cells were more con-sistent in the d16HER2 model, indicating that only d16HER2-driven tumor growth and aggressiveness remain highly depen-dent on oncogenic signaling pathways directed by and down-stream of pd16HER2D. These in vivo data are consistent withimplications of an in vitro study reporting that trastuzumab ispreferentially active against tumors driven predominantly byHER2 homodimer-induced signaling (38).

The d16HER2 variant, which seems to stabilize HER2 homo-dimer expression and activation, activates multisignaling cas-cades (10, 19, 20), including consistent phosphorylation of SRCkinase (10, 19). SRC is the prototypic member of a non-RTKfamily with broadly pleiotropic effects on mammalian cells,including effects on cell morphology, adhesion, angiogenesis,migration, invasiveness, proliferation, differentiation, and sur-vival (39). Aberrant expression and activation of SRC occurs inseveral tumor types and has been correlated with poor out-come; SRC is also a potent mediator of many downstreameffects of both HER1 and HER2 (39). In addition, SRC is areportedly common node downstream of multiple resistancepathways and a driver of trastuzumab resistance, as it ishyperactivated in various trastuzumab-resistance cell models

Figure 5. Expression andcoexpression of HER2 and pSRCmarkers and association betweend16HER2 transcript and pSRCexpression and risk of relapse inhuman HER2-overexpressingbreast cancers patientstreated with trastuzumab. Aand B, representativeimmunofluorescence images ofhuman breast cancer tissues wereevaluated by confocal microscopyand classified according to low (A)and high (B) pSRC scores. pSRC(green) and HER2 (red) stainingindicate breast cancer cells. Nucleiwere counterstained with DRAQ5(blue). C, association betweend16HER2 transcript levelsmeasured by qRT-PCR and pSRC(%) expression in 43 human HER2-overexpressing breast cancers.�, P ¼ 0.0482 by unpaired t test. D,Pearson correlation betweenpSRC (%) and d16HER2 transcriptexpression levels in 18 qRT-PCR-tested cases positive for pSRC(>0). E, association between pSRClevels (low, <20%; high, �20%)with relapse-free survival in 84HER2-positive breast cancerpatients treated with trastuzumab.

Castagnoli et al.





(40). While Mitra and colleagues (10) speculated that ind16HER2-positive transfectants, SRC kinase might act as a"master regulator" of the spliced isoform, stabilizing its expres-sion and coupling to mitogenic and cell motility pathways andcontributing to trastuzumab resistance, we found that highlevels of pd16HER2D and M, and not pWTHER2, were directlylinked to marked SRC activity and that in vivo d16HER2-driventumorigenicity was significantly halted by trastuzumab treat-ment.We also found a consistent decrease in SRC activity uponknockdown of activated d16HER2D and Mwith the anti-HER2MAb MGR2 and with MAb 4D5, the murine precursor oftrastuzumab, strongly suggesting that the pd16HER2D–pSRCsignaling axis is particularly sensitive to trastuzumab admin-istration. Additional evidence for a functional cross-talkbetween pd16HER2D and pSRC came from both IHC andconfocal microscopy analyses, indicating that such moleculesare coexpressed at high levels at the cell surface by the sametumor cells in either primary mammary lesions or lung metas-tases of the d16HER2 tg line. Overall, our preclinical findingssuggest that intense coexpression of the d16HER2 variant andpSRC at the tumor cell membrane reflects pd16HER2D-drivensignaling.In light of our previous speculation that the proportion and

relevance of d16HER2 in HER2-positive breast cancers mightredefine its role in sensitivity/resistance to trastuzumab andcan have an impact on current therapeutic strategies (32), wesought clinical verification of our preclinical data by examiningtissue from 84 HER2-positive breast cancers treated withadjuvant trastuzumab (41). In 43 out of 84 breast cancerspecimens for which frozen samples were available ford16HER2qRT-PCR analysis, 12 out of 13 high-pSRC–expressing

primary tumors expressed elevated levels of d16HER2 tran-script, strongly suggesting that pSRC reflects activatedd16HER2 homodimers in human HER2-positive breast can-cers. Indeed, such tumors are enriched in "tumor metastasis,""hypoxia," and "cell motility" pathways, all features of aggres-siveness revealed in the d16HER2 tg model. Thus, the betterprognosis observed in the trastuzumab-treated HER2-positivebreast cancer patients with elevated pSRC could be a directconsequence of the expression on their tumors of an activatedd16HER2–SRC signaling axis, as observed in the trastuzumab-sensitive d16HER2-driven mouse model. Indeed, in silico anal-yses to better define the high-pd16HER2/pSRC tumor profileindicated an "activated-d16HER2 metagene" that wasexpressed at significantly higher levels in tumors completelyresponsive to neoadjuvant trastuzumab-based therapy as com-pared with those only partially responsive, whereas "activated-d16HER2 metagene" expression levels did not differ betweencomplete and partial responders to neoadjuvant chemother-apy alone.

Our findings seem to contrast directly with those of Zhangand colleagues (40), who reported a lower clinical response rateand a higher progressive disease rate after trastuzumab treat-ment in HER2-positive breast cancer patients with high pSRCexpression; however, it should be noted that their seriesconsisted of 57 breast cancer patients who received first-linetrastuzumab-based therapy in a metastatic setting, whereasour series includes breast cancer patients treated with tras-tuzumab-based regimens in an adjuvant setting (41). Wespeculate that although HER2-positive primary breast cancersexpressing high levels of pSRC are initially dependent on HER2and all its potential driver isoforms and are, thus, responsive to

Figure 6. A, ROC curve forexpression of "activated-d16HER2metagene" in d16HER2-high,pSRC-high tumors. B, associationbetween "activated-d16HER2metagene" expression andresponse to trastuzumab-basedneoadjuvant therapy in theGSE22358 dataset. �, P ¼ 0.0305by unpaired t test. C, ROC curve oftrastuzumab response predictionusing "activated-d16HER2metagene" in the GSE22358dataset. D, association between"activated-d16HER2 metagene"expression and response toneoadjuvant chemotherapy in theGSE41656 dataset. CR, completeresponse; PR, partial response.






trastuzumab, the progression of such breast cancers due to ahigh HER2-dependent growth ratemight lead to accumulationof genetic alterations that result in less HER2 dependency and,in turn, significantly less or even no responsiveness totrastuzumab.

Such a hypothesis would reconcile the contrasting findingsof Zhang and colleagues (40) with our own, as HER2 signaling–dependent tumors benefitting from trastuzumab at an earlystagemay just be those that, if not treated early, gain additionaldependencies in a metastatic setting to allow escape fromtrastuzumab therapeutic effects. Together, our findings indi-cate the need for further evaluation of the role of pSRC inprimary and advanced HER2-positive disease before clinicaldecision making. While the relatively small size of our HER2patient samples precluded analysis of whether breast cancerpatients with high d16HER2/low pSRC levels might expressinactive d16HER2 homodimers due to a failure to couple withpSRC, and despite a lack of a specific anti-pd16HER2D reagent,our present preclinical, clinical, and in silico data support thenotion that activated HER2 signaling is indicative of benefitsfrom the addition of trastuzumab to chemotherapy and thatd16HER2 expression is not a reliable indicator of trastuzumabresistance but instead a mirror of pSRC activity, reflectingd16HER2 homodimer-mediated driver activity leading to highresponsiveness to trastuzumab. These datamight shed light onthe very complex "HER2 world" and help clinicians identifyingthe "real" HER2 drivers for targeting by appropriate pharma-cologic strategies.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: L. Castagnoli, M. Iezzi, P.L. Lollini, P. Nanni,E. Tagliabue, S.M. PupaDevelopment ofmethodology: L. Castagnoli, M. Iezzi, V. Ciravolo, G.Marzano,R.A. Lamolinara, R. Zappasodi, P. Gasparini, A. PalladiniAcquisition of data (provided animals, acquired and managed pati-ents, provided facilities, etc.): L. Castagnoli, M. Iezzi, A. Lamolinara,P. Gasparini, M. Campiglio, A. Amici, C. Chiodoni, A. Palladini, P.L. Lollini,T. Triulzi, P. Nanni, S.M. PupaAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): L. Castagnoli, M. Iezzi, G.C. Ghedini, V. Ciravolo,A. Lamolinara, P. Gasparini, C. Chiodoni, A. Palladini, P. Nanni, T. Triulzi,E. Tagliabue, S.M. PupaWriting, review, and/or revision of the manuscript: L. Castagnoli, M. Iezzi,P.L. Lollini, S. Menard, P. Nanni, E. Tagliabue, S.M. PupaAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): L. Castagnoli, G.C. Ghedini, V. Ciravolo,A. Lamolinara, M. Campiglio, A. Palladini, T. Triulzi, P. Nanni, E. Tagliabue, S.M.PupaStudy supervision: E. Tagliabue, S.M. Pupa

AcknowledgmentsThe authors thank Piera Aiello and Cristina Ghirelli for technical assistance,

the personnel of Tissue Bank of the Fondazione IRCCS Istituto Nazionale deiTumori, Milan, for providing breast cancer frozen samples, and LauraMameli forsecretarial assistance. The authors also thank Drs. M.P. Colombo, G. Sozzi, andM. Di Nicola for critical reading of the article.

Grant SupportThis work was funded by grants from the Associazione Italiana Ricerca

Cancro (AIRC) 10352 to S.M. Pupa; Fellowship 15002 to L. Castagnoli; Fellowship12595 to G.C. Ghedini; and Ministero Italiano della Salute RF-2009-1532281 toS.M. Pupa.

The costs of publication of this article were defrayed in part by the payment ofpage charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received April 4, 2014; revised July 4, 2014; accepted August 11, 2014;published OnlineFirst August 27, 2014.

References1. Yarden Y, Sliwkowski MX. Untangling the ErbB signalling network. Nat

Rev Mol Cell Biol 2001;2:127–37.2. ArteagaCL,SliwkowskiMX,OsborneCK,Perez EA, Puglisi F, Gianni L.

Treatment of HER2-positive breast cancer: current status and futureperspectives. Nat Rev Clin Oncol 2012;9:16–32.

3. Harris CA, Ward RL, Dobbins TA, Drew AK, Pearson S. The efficacy ofHER2-targeted agents in metastatic breast cancer: a meta-analysis.Ann Oncol 2011;22:1308–17.

4. Tagliabue E, Campiglio M, Pupa SM,M�enard S, Balsari A. Activity andresistance mechanisms of trastuzumab in different clinical settings.Cancer Treat Rev 2012;38:212–7.

5. Scaltriti M, Rojo F, Ocana A, Anido J, Guzman M, Cortes J, et al.Expression of p95HER2, a truncated form of the HER2 receptor, andresponse to anti-HER2 therapies in breast cancer. J Natl Cancer Inst2007;99:628–38.

6. Pedersen K, Angelini PD, Laos S, Bach-Faig A, Cunningham MP,Ferrer-Ramon C, et al. A naturally occurring HER2 carboxy-terminalfragment promotes mammary tumor growth and metastasis. Mol CellBiol 2009;29:3319–31.

7. Bose R, Kavuri SM, Searleman AC, ShenW, Shen D, Koboldt DC, et al.ActivatingHER2mutations inHER2gene amplification negative breastcancer. Cancer Discov 2013;3:224–37.

8. Kwong KY, Hung MC. A novel splice variant of HER2 with increasedtransformation activity. Mol Carcinog 1998;23:62–8.

9. Castiglioni F, TagliabueE,CampiglioM,PupaSM,Balsari A,M�enardS.Role of exon-16-deleted HER2 in breast carcinomas. Endocr RelatCancer 2006;13:221–32.

10. Mitra D, Brumlik MJ, Okamgba SU, Zhu Y, Duplessis TT, Parvani JG,et al. An oncogenic isoform of HER2 associated with locally dissem-

inated breast cancer and trastuzumab resistance. Mol Cancer Ther2009;8:2152–62.

11. Muller WJ, Sinn E, Pattengale PK, Wallace R, Leder P. Single-stepinduction ofmammary adenocarcinoma in transgenicmice bearing theactivated c-neu oncogene. Cell 1988;54:105–15.

12. Guy CT, Webster MA, Schaller M, Parsons TJ, Cardiff RD, Muller WJ.Expression of the neu protooncogene in the mammary epithelium oftransgenic mice induces metastatic disease. Proc Natl Acad Sci U S A1992;89:10578–82.

13. BargmannCI,Weinberg RA. Oncogenic activation of the neu-encodedreceptor protein by point mutation and deletion. EMBO J 1988;7:2043–52.

14. Bouchard L, Lamarre L, Tremblay PJ, Jolicoeur P. Stochastic appear-ance of mammary tumors in transgenic mice carrying the mmtv/c-neuoncogene. Cell 1989;57:931–6.

15. Suda Y, Aizawa S, Furuta Y, Yagi T, Ikawa Y, Saitoh K, et al.Induction of a variety of tumors by c-erbB2 and clonal nature oflymphomas even with the mutated gene (Val659-Glu659). EMBO J1990;9:181–90.

16. Piechocki MP, Ho YS, Pilon S, Wei WZ. Human ErbB-2 (Her-2)transgenic mice: a model system for testing Her-2 based vaccines.J Immunol 2003;171:5787–94.

17. Finkle D, Quan ZR, Asghari V, Kloss J, Ghaboosi N, Mai E, et al. HER2-targeted therapy reduces incidence and progression of midlife mam-mary tumors in female murine mammary tumor virus huHER2-trans-genic mice. Clin Cancer Res 2004;10:2499–511.

18. Ursini-Siegel J, Schade B, Cardiff RD, Muller WJ. Insights from trans-genic mouse models of ERBB2-induced breast cancer. Nat RevCancer 2007;7:389–97.

Castagnoli et al.





19. Marchini C, Gabrielli F, Iezzi M, Zanobi S, Montani M, Pietrella L, et al.The human splice variant delta16HER2 induces rapid tumor onset in areporter transgenic mouse. PLoS ONE 2011;6:e18727.

20. Alajati A, Sausgruber N, Aceto N, Duss S, Sarret S, Voshol H, et al.Mammary tumor formation and metastasis evoked by a HER2 splicevariant. Cancer Res 2013;73:5320–7.

21. MagnificoA, Albano L, Campaner S, Delia D, Castiglioni F,Gasparini P,et al. Tumor-initiating cells of HER2-positive carcinoma cell linesexpress the highest oncoprotein levels and are sensitive to Trastuzu-mab. Clin Cancer Res 2009;15:2010–21.

22. Ghedini GC, Ciravolo V, Tortoreto M, Giuffre S, Bianchi F, CampiglioM, et al. Shed HER2 extracellular domain in HER2-mediated tumorgrowth and in trastuzumab susceptibility. J Cell Physiol 2010;225:256–65.

23. Wickenden JA, Jin H, Johnson M, Gillings AS, Newson C, Austin M,et al. Colorectal cancer cells with the BRAF(V600E) mutation areaddicted to the ERK1/2 pathway for growth factor-independent sur-vival and repression of BIM. Oncogene 2008;27:7150–61.

24. Triulzi T, Casalini P, Sandri M, Ratti F, Carcangiu ML, Colombo MP,et al. Neoplastic and stromal cells contribute to an extracellular matrixgene expression profile defining a breast cancer subtype likely toprogress. PLoS ONE 2013;8:e56761.

25. Cox DR. Regression models and life tables. J R Stat Soc 1972;34:187–220.

26. SubramanianA, TamayoP,Mootha VK,Mukherjee S, Ebert BL,GilletteMA, et al. Gene set enrichment analysis: a knowledge-based approachfor interpreting genome-wide expression profiles. Proc Natl Acad SciU S A 2005;102:15545–50.

27. Gluck S, Ross JS, RoyceM,McKenna EF Jr, Perou CM, Avisar E, et al.TP53 genomics predict higher clinical and pathologic tumor responsein operable early-stage breast cancer treated with docetaxel-capeci-tabine þ/� trastuzumab. Breast Cancer Res Treat 2012;132:781–91.

28. de Ronde JJ, Rigaill G, Rottenberg S, Rodenhuis S, Wessels LF.Identifying subgroup markers in heterogeneous populations. NucleicAcids Res 2013;41:e200.

29. Hajian-Tilaki KO, Hanley JA, Joseph L, Collet JP. A comparison ofparametric and nonparametric approaches to ROC analysis of quan-titative diagnostic tests. Med Decis Making 1997;17:94–102.

30. Goldhirsch A, Wood WC, Coates AS, Gelber RD, Thurlimann B, SennHJ. Strategies for subtypes–dealingwith the diversity of breast cancer:

highlights of the St. Gallen International Expert Consensus on thePrimary Therapy of Early Breast Cancer 2011. Ann Oncol 2011;22:1736–47.

31. Shiu KK, Wetterskog D, Mackay A, Natrajan R, Lambros M, Sims D,et al. Integrativemolecular and functional profiling of ERBB2-amplifiedbreast cancers identifies new genetic dependencies. Oncogene2014;33:619–31.

32. Jackson C, Browell D, Gautrey H, Tyson-Capper A. Clinical signifi-cance of HER-2 splice variants in breast cancer progression and drugresistance. Int J Cell Biol 2013;2013:973584.

33. Zagozdzon R, Gallagher WM, Crown J. Truncated HER2: implica-tions for HER2-targeted therapeutics. Drug Discov Today 2011;16:810–6.

34. Krop IE, Burstein HJ. Trastuzumab: qui bono? J Natl Cancer Inst2013;105:1772–5.

35. Pupa SM, Campiglio M, Rossini A, Orlandi R, Ciravolo V, Amici A,et al., Role of D16HER2 splice variant in breast tumor progressionand response to HER2-targeted therapy. In: Williams SI, Rogers CE,editors. HER2 and Cancer: Mechanism, Testing and TargetedTherapy. Hauppauge, NY: Nova Science Publishers, Inc. 2011. p.201–9.

36. Garrett JT, Arteaga CL. Resistance to HER2-directed antibodies andtyrosine kinase inhibitors: mechanisms and clinical implications. Can-cer Biol Ther 2011;11:793–800.

37. Rexer BN, Arteaga CL. Intrinsic and acquired resistance to HER2-targeted therapies in HER2 gene-amplified breast cancer: mechan-isms and clinical implications. Crit Rev Oncog 2012;17:1–16.

38. GhoshR,NarasannaA,WangSE, Liu S,ChakrabartyA, Balko JM, et al.Trastuzumab has preferential activity against breast cancers driven byHER2 homodimers. Cancer Res 2011;71:1871–82.

39. Guarino M. Src signaling in cancer invasion. J Cell Physiol 2010;223:14–26.

40. Zhang S, Huang WC, Li P, Guo H, Poh SB, Brady SW, et al.Combating trastuzumab resistance by targeting SRC, a commonnode downstream of multiple resistance pathways. Nat Med 2011;17:461–9.

41. Campiglio M, Bufalino R, Sasso M, Ferri E, Casalini P, Adamo V, et al.Effect of adjuvant trastuzumab treatment in conventional clinicalsetting: anobservational retrospectivemulticenter Italian study.BreastCancer Res Treat 2013;141:101–10.






2014;74:6248-6259. Published OnlineFirst August 27, 2014.Cancer Res Lorenzo Castagnoli, Manuela Iezzi, Gaia C. Ghedini, et al. Efficacy for TrastuzumabActivated d16HER2 Homodimers and SRC Kinase Mediate Optimal

Updated version

10.1158/0008-5472.CAN-14-0983doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2014/08/26/0008-5472.CAN-14-0983.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/74/21/6248.full#ref-list-1

This article cites 40 articles, 11 of which you can access for free 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://cancerres.aacrjournals.org/content/74/21/6248To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-14-0983

http://cancerres.aacrjournals.org/content/suppl/2014/08/26/0008-5472.CAN-14-0983.DC1

http://cancerres.aacrjournals.org/content/74/21/6248.full#ref-list-1

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

mailto:[email protected]

http://cancerres.aacrjournals.org/content/74/21/6248


activated d16her2 homodimers and src kinase mediate … · therapeutics, targets, and chemical...

Documents