antibody targeting of cell-bound muc1 sea domain kills ... · antibodies was done with horseradish...

Therapeutics, Targets, and Chemical Biology

Antibody Targeting of Cell-Bound MUC1 SEA Domain KillsTumor Cells

Edward Pichinuk1, Itai Benhar2, Oded Jacobi1, Michael Chalik1, LotemWeiss1, Ravit Ziv1, Carolyn Sympson3,Amolkumar Karwa3, Nechama I. Smorodinsky1, Daniel B. Rubinstein4, and Daniel H. Wreschner1,5

AbstractThe cell-surface glycoprotein MUC1 is a particularly appealing target for antibody targeting, being

selectively overexpressed in many types of cancers and a high proportion of cancer stem–like cells. Howeverthe occurrence of MUC1 cleavage, which leads to the release of the extracellular a subunit into the circulationwhere it can sequester many anti-MUC1 antibodies, renders the target problematic to some degree. Toaddress this issue, we generated a set of unique MUC1 monoclonal antibodies that target a region termed theSEA domain that remains tethered to the cell surface after MUC1 cleavage. In breast cancer cell populations,these antibodies bound the cancer cells with high picomolar affinity. Starting with a partially humanizedantibody, DMB5F3, we created a recombinant chimeric antibody that bound a panel of MUC1þ cancer cellswith higher affinities relative to cetuximab (anti-EGFR1) or tratuzumab (anti-erbB2) control antibodies.DMB5F3 internalization from the cell surface occurred in an efficient temperature-dependent manner.Linkage to toxin rendered these DMB5F3 antibodies to be cytotoxic against MUC1þ cancer cells at lowpicomolar concentrations. Our findings show that high-affinity antibodies to cell-bound MUC1 SEA domainexert specific cytotoxicity against cancer cells, and they point to the SEA domain as a potential immunogen togenerate MUC1 vaccines. Cancer Res; 72(13); 3324–36. �2012 AACR.

IntroductionMUC1 is a mucin-like glycoprotein that can generate a

variety of differing isoforms (1). Of these, the most intenselystudied has been a polymorphic type I high molecular weighttransmembrane protein (MUC-TM) consisting of an extracel-lular domain containing 20 to 125 tandem repeats of 20 aminoacids, followed by a transmembrane domain and a shortcytoplasmic tail (2–4). MUC1 is a heterodimer that is cleavedsoon after synthesis within the SEAmodule, a highly conserveddomain of 120 amino acids (4–6). Cleavage of MUC1 yields 2unequal chains: a large extracellular a subunit containing thetandem repeat array specifically bound in a strong noncovalentinteraction to a smaller b subunit containing the transmem-brane and cytoplasmic domains of the molecule (4, 7).

MUC1 is highly expressed on a range of malignancies,including breast, pancreas, ovarian, prostate, and colon car-cinomas, as well as on the malignant plasma cell of myeloma(8–13). Because of this overexpression, MUC1 has been the

subject of a great amount of attention, primarily for itspotential as a target for tumor-specific therapies. In fact, basedon a number of ranked criteria required of an optimal cancervaccine candidate, theMUC1proteinwas listed by theNationalCancer Institute Pilot Project the second best target from a listof 75 potential tumor-associated antigens (14). Moreover,within the categories of expression level, stem cell expressionand number of patients with antigen-positive cancers, theMUC1 protein received perfect scores (14).

Most anti-MUC1 antibodies reported to date target thehighly immunogenic tandem repeat array of the MUC1 achain (for e.g., refs. 15–17). Because the MUC1 a chain is notdirectly tethered to the cell surface, it is also found in theperipheral circulation. There it sequesters circulating anti-tandem repeat antibodies, limiting their ability to reach theMUC1þ tumor cells (15, 16). In addition, deposition ofimmune complexes of antitandem a chain antibodies andcirculating a chain potentially may result in end-organdamage (16).

Targeting MUC1 epitopes bound to the cell surface wouldavoid antibody sequestration by circulating a chain. TheMUC1 SEA domain formed by the interaction of the a subunitwith the extracellular portion of b subunit is an intricatestructure, which remains fixed to the cell surface (4, 6) andsignificantly, results in a stable target structure. In previousinitial studies, we showed the proof-of-principle of generatingantibodies that specifically recognize the MUC1 SEA domain(18). The prototype DMC209 mAb described (18) has 2 dis-advantages that compromise its potential clinical use: It is anIgM and it has relatively low affinity for its target.

Authors' Affiliations:Departments of 1Cell Research and Immunology and2Molecular Microbiology and Biotechnology, Tel Aviv University, RamatAviv, Israel; 3Biological Sciences, Covidien Pharmaceuticals, St. Louis,Missouri; 4National Cancer Institute, NIH, Bethesda, Maryland; and 5Bio-modifying, Inc., San Diego, California

Corresponding Author: Daniel H. Wreschner, Department ofCell Research and Immunology, University of Tel Aviv, Ramat Aviv,Israel 69978. Phone: 972-3-640-7425; Fax: 972-3-642-2046; E-mail:[email protected]

doi: 10.1158/0008-5472.CAN-12-0067

�2012 American Association for Cancer Research.

CancerResearch

Cancer Res; 72(13) July 1, 20123324

on June 30, 2020. © 2012 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst April 16, 2012; DOI: 10.1158/0008-5472.CAN-12-0067

http://cancerres.aacrjournals.org/

As a result of these considerations, we sought highly specific,anti-MUC1 antibodies directed against the a/b junction thatwere also affinity maturated and not of the IgM subclass. Wereport here the generation of 7 novel monoclonal antibodies(mAb), 5 Ig-g1, and 2 IgA that bind theMUC1 protein within itsSEA domain with picomolar affinities. Our findings suggestthat the IgG1 and IgA mAbs targeting the MUC1 SEA domainmay prove beneficial in the therapy of MUC1þ malignancieswhen either administered alone or in conjunction with otherantitumor antibodies and chemocytotoxic agents.

Materials and MethodsMaterials and antibodiesReagents and chemicals were obtained from Sigma, unless

otherwise specified. Anti-MUC1 SEA-module mAbs were gen-erated as described below. The anti-MUC1 tandem repeatantibodies used here were as previously described (19, 20).

Cell linesHuman breast carcinoma cell lines T47D, MCF7, and ZR75,

human epidermoid carcinoma A431 and KB cell lines, humangastric carcinoma N87 cell line, and human pancreatic carci-noma Colo357 cell line were maintained as previouslydescribed (18).

Cell cultureCells were grown at 37�C and 5% CO2, in culture media

supplemented with 10% heat-inactivated fetal calf serum(FCS), 2 mmol/L L-glutamine, 100 IU/mL penicillin, and25 mg/mL streptomycin. DA3 (mouse mammary tumor cells)and CHO-KI (Chinese Hamster Ovary cells) were grown inDulbecco's Modified Eagle's Medium (DMEM) or DMEM: F12nutrient mixture (1:1), respectively.

Generation of stable DA3 mouse mammary tumor celltransfectants expressing MUC1-TMDA3 cells were cotransfected with the eukaryotic expression

plasmids pCL-MUC1-TM and with pSV2neo (which codes forneomycin resistance). Expression constructs were transfectedinto cells using the calcium phosphate procedure. Stabletransfectants were selected with neomycin.

Generation of bacterial recombinant MUC1-X,SEA, SEA-4G proteinsBacterial MUC1-X, SEA, SEA-4G proteins were prepared as

the (His)6-MUC1-Xex, (His)6-SEA, and (His)6-SEA-4G fusionproteins as previously described (21) and comprise 6 histidineresidues at their N-terminus, followed by the extracellulardomain of the MUC1-X/SEA/SEA-4G proteins, respectively.

Immunization of mice and generation of hybridomasMice were initially immunized with 5 consecutive intrader-

mal DNA immunizations spaced at 21-day intervals. Theimmunizing DNA consisted of the pCL-MUC1-TM expressionvector plasmid that codes for the MUC1-TM protein (21). Theextracellular domain of MUC1-X protein (recombinant bacte-rial MUC1-Xex, see below) together with incomplete Freund's

adjuvant was then used to boost the mice. Bacterially synthe-sized recombinant MUC1-Xex protein used for these immu-nizations spontaneously self-cleaves as previously described(4), generating the MUC1-X a and b subunits that strongly, yetnoncovalently, interact with each other forming the very stableheterodimeric cleaved MUC1-Xex protein (depicted in Figs. 1and 4). Hybridomas were prepared by fusion of nonsecretingmyeloma cells with immune splenocytes and screened byELISA assay (see below).

ELISA for determining binding of anti-MUC1 polyclonaland monoclonal antibodies to the extracellular domainof the MUC1-X protein

Elisa Immunoassay plates (CoStar) were coated with re-combinant MUC1 proteins followed by blocking. Spent culturemedia from the initial hybridomas was then applied tothe wells. Following incubation, samples were removed andthe wells were washed with PBS/Tween. Detection of boundantibodies was done with horseradish peroxidase (HRP)-con-jugated anti-mouse antibody.

Two-tiered screening for selection of anti-MUC1monoclonal antibodies

The primary screen of the hybridomas was carried out byassessing antibody binding to the extracellular domain of

C

MUC1/TM

MUC1/X

[cleaved]

C

N

α-subunit

β-subunit

N30

N30

VN

TR

N

Intracellular

Extracellular

Anti-α/β junction

antibodies

mAbsDMB4B4 Igγ1

DMB4F4 Igγ1

DMB5F3 Igγ1

DMB7F3 Igγ1

DMB10F10 Igγ1

DMB10B7 IgA

DMB13D11 IgA

DMC209 IgM

Figure 1. Scheme of MUC1 proteins used for immunizing mice andsubsequent screening for anti-MUC1 SEA module a/b junction mAbs.Schema of isoforms structures. The cleaved full-length MUC-TM proteincontaining both the b-chain and the a-chain with its tandem repeat arrayand the cleaved MUC1-X protein. The domains are color schemed asfollows: signal peptide, purple; N-terminal 30 amino acids (N30), yellow;sequences flanking the tandem repeat array, green; central tandemrepeat array, light blue; N-terminal part of SEA domain, red;C-terminal part of SEA domain, hatched red and white; transmembranedomain, dark blue and cytoplasmic domain, pink. The SEA domaincomprising the interacting a and b subunits is indicated by the dashedblack circles. Screening of hybridoma supernatants for antibodies thatselectively recognize theSEAmodulea/b junctionwas carried outwith anELISA screen using wells precoated with MUC1-Xex protein (seeMaterials and Methods). The series of 7 anti-MUC1 antibodies (the DMBseries, red lettering), as well as the previously described (18) IgMDMC209, bind the SEA module.

Picomolar Affinity Monoclonals for Cancer Cell Eradication

www.aacrjournals.org Cancer Res; 72(13) July 1, 2012 3325




MUC1-X (MUC1-Xex) as described in the ELISA assay (above).To select hybridomas secreting antibody that recognize notonly MUC1-Xex but also the complete cell surface MUC1-TMprotein, those hybridomas presenting a positive signal in thefirst screen were subjected to a second-tier screen. This con-sisted of flow cytometric analysis using mouse cell transfec-tants (DA3-TM) expressing human MUC1-TM and, in parallel,to the same parental cells (DA3-PAR) that do not expresshuman MUC1. This procedure ensured selection of antibodythat not only bound MUC1 moieties common to both theMUC1-X and MUC1-TM but also recognized cell surfacehuman MUC1-TM as expressed by mammalian cells.

Isolation of MCF7 breast cancer side population ofcancer cells

An MCF7 breast cancer cell subpopulation that has beenreported to show some characteristics of cancer stem cells (22)were isolated as described by Finn (22). Briefly, MCF7 cellswere stained with the fluorescent dye Hoechst 33342. Due totheir increased efflux of the dye a side population ofMCF7 cellswas isolated by flow cytometry. The resultant cells were thenconfirmed as showing some characteristics of stem cells usinga battery of stem cell markers.

Preparation of ZZ-Pseudomonas exotoxin andchDMB5F3:ZZ-PE38 immunotoxin

The ZZ-PE38 fusion protein comprises the Pseudomonasexotoxin PE38 and the ZZ domain derived from protein A. TheZZ portion binds to the Fc region of IgG to form immunotoxinconjugates. Plasmid pET22-NN-ZZ-PE38 was used for expres-sion of soluble ZZ–Pseudomonas exotoxin A (PE38) fusionprotein secreted to the periplasm of BL-21 (DE3) Escherichiacoli cells, which was then purified as previously described (23).

SDS-PAGE/Western blotProteins separated on SDS-PAGE were electrotransferred

for 2 hours at 0.5 Amp on to nitrocellulose filters in transferbuffer. Blots were blocked with 5% skimmed milk followed byincubation with the primary antibody (anti-MUC1 antibodies,see Results). Bound primary antibody was detected withsecondary anti-mouse antibody conjugated to HRP (ChemiconInternational) followed by enhanced chemiluminescence.

Cell killing assayMousemammary tumor cells transfected and stably expres-

sing human MUC1-TM (DA3-TM), the parental cells that donot express human MUC1 (DA3-PAR wild type), and MUC1þ

human cell lines (T47D and ZR75 breast cancer cells, Colo357pancreatic cancer cells, as well as additional cancer cells asdetailed in Results), were grown in DMEM supplemented with10% FCS. These cells (10,000 cells per well in 100 microliters ofmedium) were seeded in 96-well cell culture plates (Corning)and grown at 37�C in 5% CO2 in culture media. Five hours afterseeding, 50 microliters of the chDMB5F3 was diluted bydecimal dilutions starting from 1,600 pm and mixed with 50microliters ZZ-PE38 toxin. The immunotoxin mixture wasdirectly applied to the cells (final ZZ-PE38 toxin concentrationwas 250 ng/mL). Negative controls included adding to the

target cells the following (instead of the chDMB5F3:ZZ-PE38toxin immunoconjugates)—(i) concentrated medium fromparental CHO-K1 cells that is devoid of the anti-MUC1chDMB5F3 mAb, (ii) ZZ-PE38 alone, (iii) chDMB5F3 mAbalone, devoid of the ZZ-PE38 toxin. Cell viability was assessedby measuring alkaline phosphatase activity per well.

Construction of vectors for mammalian expression inChinese hamster ovary cells of recombinant DMB-5F3–constant region of human IgG1 [C-hIgG1]

Methods used to generate these vectors essentially followedthose previously reported (23). Themammalian vectors pMAZ-IgH and pMAZ-IgL (19) were used as backbones for theexpression of the VH and VL regions of DMB5F3 fused tohuman g1 heavy and human k light chains, respectively.

Surface plasmon resonance binding assayRecombinant MUC1-X protein was immobilized on a

CM5 chip. As control, anti-CD20 IgG (Rituxan) was treated ina like manner. Remaining active groups were saturated with1 mmol/L ethanolamine. Surface plasmon resonance (SPR)was done using Biacore3000 according to manufacturer'sspecifications (GE Healthcare). Serial dilutions of antibodyDMB5F3 (0.4–7 nmol/L) were measured for 4 minutes asso-ciation and 30 minutes dissociation. The chip was regeneratedwith 5 mmol/L NaOH. Data fitting was carried out usingalgorithm of the BIAevalution software.

Preparation of DMB5F3 FabDMB5F3 IgG1 was reacted with papain to give a papain-to-

antibody ratio of 1:20 (ww) followed by incubation at 37�C for 5hours. The reaction was stopped with crystalline iodoaceta-mide at a final concentration of 0.03 mol/L. The resultantmixture was dialyzed against PBS pH8, passed on a protein Gcolumn, and Fab fractions collected.

Analysis of binding to MUC1/Her2/EGFR1-expressingtumor cells by flow cytometry

Evaluation of binding by flow cytometry was carried out asfollows: 5 � 105 cells were used in each experiment. Aftertrypsinization, cells were washed and mAbs at varying dilu-tions were added to the cell tubes for 1 hour at 4�C. Afterwashingwithfluorescence-activated cell sorting (FACS) buffer,fluorescein isothiocyanate-labeled goat anti-human Fc anti-bodywas added to the tubes for 45minutes at 4�C. Detection ofbound IgG was done by means of flow cytometry on a FACS-Calibur (Becton Dickinson), and results analyzed with the-CELLQuest program (Becton Dickinson).

ResultsGeneration of monoclonal antibodies to the cell-boundMUC1 a/b junction

To avoid generating antibodies to the circulating a chain,theMUC1 immunogen should ideally be cell bound at all times.To comply with these criteria, we used the extracellulardomain of the naturally occurring MUC1-X isoform thatincorporates elements derived from both the extracellular a

Pichinuk et al.

Cancer Res; 72(13) July 1, 2012 Cancer Research3326




chain and the cell bound b chain. This provided the advantageof incorporating the cell-bound region consisting of the a/bjunction present in the full naturally occurring MUC1 protein(Fig. 1), yet at the same time is devoid of the highly immuno-genic tandem repeat array (Fig. 1,MUC1-X). TheMUC1 proteinused for these immunizations is cleaved MUC1-Xex proteinsynthesized in bacteria. Within the bacteria, MUC1-Xex under-goes spontaneous self-cleavage and comprises the noncova-lently interacting MUC1 a and b subunits.Using our immunization protocol (Materials and Methods),

high titer anti-MUC1-X polyclonal antibody sera up to1:100,000 dilutions were obtained. Spleens of such mice wereused for hybridoma formation and hybridoma supernatantssubjected to a 2-tiered screen consisting of (i) binding toMUC1-Xprotein, and (ii) amore stringent second tier assessing

binding to MUC1-expressing cancer cells as assayed by flowcytometry. Hybridoma formation resulted in 7 mAbs, 5 of theIg-g1 isotype, designated DMB4B4, DMB4F4, DMB5F3,DMB7F3, and DMB10F10. Surprisingly 2 monoclonals of theIgA subclass were also isolated, DMB10B7 and DMB13D11.

Cytometric analyses of monoclonal antijunctionalantibodies

Flow cytometric analyses showed that all 7 mAbs boundstrongly to DA3-TM cells that express the full-length MUC1. Incontrast, untransfected DA3-PAR cells that do not expressMUC1 were consistently negative with all antibodies (Fig. 2).Similarly, Colo357, a MUC1þ pancreatic cancer cell lineshowed unequivocal reactivity with all anti-MUC1 monoclo-nals (Fig. 2) as did several breast cancer cell lines, such as T47D,

DA3-PAR

DMB7F3

Ig-γ1

DMB4F4

Ig-γ1

DMB10F10

Ig-γ1

DMB4B4

Ig-γ1

DMB5F3

Ig-γ1

DMB13D11

IgA

DMB10B7

IgA

DA3-MUC1 COLO357 COLO357 + MUC1-Xex

Figure 2. Flow cytometric analyses of DMB anti-MUC1 SEA module a/b junction mAbs with MUC1-TM expressing cells. Parental nontransfected mousemammary tumor DA3 cells and the same cells transfected with cDNA coding for and expressing full-length MUC1-TM. Columns designated DA3-PAR andDA3-MUC1 were reacted with the DMB series of anti-MUC1 SEA module a/b junction mAbs. Flow cytometry with the fluorescently labeled secondaryantibody alone (red tracing) and with anti-MUC1 antibody followed by secondary antibody (green tracing) are as shown. Colo357 cells were similarly reactedwith the 7mAbs, either in the absence or presence of competing soluble extracellular domain of theMUC1-Xex protein [columnColo357 and columnColo357þ MUC1-Xex].






MCF7, and ZR75 (data not shown). Further confirmation ofbinding specificity was provided by addition of soluble exog-enous MUC1-Xex protein, the extracellular domain of theMUC1-X protein, which abolished all reactivity (Fig. 2). Thesestudies underscore the fact that the DMB antibodies bind acell-bound domain noncompetitive with moieties present inthe shed a chain of the MUC1-TM protein. In addition to theflow cytometric studies, initial immunohistochemical analyseswere carried out with one of these mAbs (DMB5F3) usingsections of breast cancer tissue. These stainings showed sig-nificant DMB5F3 immunoreactivity with the breast cancercells (data not shown) results clearly, in line with the well-known overexpression of MUC1 by breast (and other) adeno-carcinoma cells comprehensively documented in numerousstudies.

Binding of antijunctional antibodies to MCF7 sidepopulation cells that show some characteristics of breastcancer stem cells

To determine whether the DMB series of anti-MUC1 junc-tional antibodies bind not only to differentiatedMUC1þ tumorcells but also to MCF7 side population cells that show some

characteristics of cancer stem cells (22), the antibodies werereacted with the MCF7 side population. Both antibodiesDMB4B4 and DMB4F4 anti-MUC1 yielded sharp shifts in thegated cell population of both the MCF7 side populationpreviously shown to bear some characteristics of cancer stemcells (ref. 22; Fig. 3, orange,middle and right panels) andmatureMCF7 cancer cells (Fig. 3, green, middle and right panels). Theside population cells bound to the anti-MUC1 antibodies wereshown to have multiple characteristics of stem/progenitorcells (22). These include (i) preferential efflux of the fluorescentDNA-binding dye Hoechst 3342, (ii) phenotypic characteriza-tion as CD44þ/CD24�/low, luminal and epithelial markersCK15 and EpCAM, and the stem cell/progenitor marker CK19,(iii) preferential formation of mammospheres in suspensionculture, and (iv) gene array including preferential expressionof a variety of protein products including Wiskott–Aldrichsyndrome interacting protein and insulin-like growth factor–binding protein. These findings indicated that the sidepopulation of MCF7 cells representing cells that show somecharacteristics of cancer stem cells (22) can be targeted byanti-MUC1 junctional antibodies DMB4B4 and DMB4F4 noless effectively than MCF7 "mature" MUC1þ tumor cells.

0% 5.1%

0.2% 94.7% 98.4%

0.3% 94.5% 98.9%

gated on SP

gated on nSP

DMB-4B4 DMB-4F4isotype

Figure 3. Binding of antijunctional antibodies to MCF7 side population cells. An MCF7 breast cancer cell population showing some characteristics of cancerstem cells was isolated as described by Finn (22). MCF7 cells were stained with the fluorescent dye Hoechst 33342. Due to their increased efflux ofthedye, a cone-shaped sidepopulationofMCF7 stemcells (5.1%of theoverall population)was isolated andcharacterized (top right plots). Reactionwith IgG1isotype control resulted in no cell shift (left bottom). Reaction with 10 mg/mL of either DMB-4B4 or DMB-4F4 resulted in near-total shifts in both theMCF7 sidecell side population (orange color, middle, and right bottom) as well as in "mature" MCF7 breast cancer cells (green color; middle and left bottom). SP, sidepopulation.

Pichinuk et al.





Despite our characterization of the side population asdescribed above, the MCF7 side population is itself a nonho-mogeneous population of cells (22), and the phenotypic iden-tification of true stem cells especially in ERþ cancer cells (suchas MCF7) remains a challenge (24, 25). Indeed the caveatshould be stressed that it is not clear whether this MCF7 sidepopulation represent stem cells as appearing in ERþ cancers.Thus to what degree the anti-MUC1 junctional antibodiestarget true stem cells will require a more precise characteri-zation of SC in ERþ tumor cells.

The DMB mAbs all recognize the SEA domain yet aredifferent from one anotherThe DMB mAbs were assessed by ELISA for binding to the

MUC1-Xex protein, the SEA domain itself, and the SEA-4Gprotein, a mutant construct consisting of uncleaved a�bbound by a 4-glycine peptide (6). These proteins were gener-ated in bacteria, and as previously reported (4), the MUC1-Xex

and SEA domain proteins spontaneously self-cleave in bacteriaat the cleavage site FRPGjSVVV, in which j indicates cleavage,generating the interacting a and b subunits (depicted inFig. 4B, bottomschema). Themutant SEA-4Gprotein (depictedin Fig. 4, bottom schema) comprises the amino acid sequenceFRPGGGGSVVV (instead of the above wild-type sequence),resulting in a noncleaved protein as previously described(6). Results showed that all 7 mAbs bound the 3 proteins (eachbound to wells of an ELISA plate) at picomolar antibodyconcentrations and confirmed that the primary target of theseantibodies is the SEA domain. Inspection of the binding curvesof the Ig-g1 mAbs to these 3 target proteins showed distinctbinding patterns for mAbs DMB5F3, DMB4F4, and DMB7F3to each of these proteins (Fig. 4A), indicating that the preciseepitope within the SEA domain is different for each of theseantibodies. Additional analyses including flow cytometry witha battery of MUC1-expressing cancer cells (data not shown)aswell asWestern blotting analyses (see below) further showed

DMB-5F3 DMB-4F4 DMB-7F3

50

100

10 1 0.010.1

Antibody concentration (nmol/L)

50

100

10 1 0.010.1


50

100

10 1 0.010.1


A

11 2

DMB 4B4DMB 4F4DMB 5F3BOS10D2

BOS10D2

α

β

1 2 1 2 1 2 1 2

β

α

α−4G−β α−4G−β

β

α−4G−β

α α

βα

βα

α−4G−β

β

SEA

module

MUC1-Xex 11 22

a. b. c. d. e.

BOS10D2

α−4G−β

SEA SEA-4G

10

15

B

Figure 4. Immunoreactivity of DMB mAbs with MUC1-Xex, SEA module, and MUC1 4G-SEA module proteins. A, ELISA plates were coated with 1 of 3recombinant bacterial MUC1 proteins extracellular domain of the MUC1-X protein (dotted lines), MUC1 SEA module (dashed lines), or the uncleaved MUC14G-SEAmodule protein (dashed and dotted lines). The indicatedmAbswere applied and bound antibodywas detected as described inMaterials andMethods.The 3 mAbs DMB-5F3, DMB-4F4, and DMB-7F3 show different patterns of binding for the 3 MUC1 proteins. B, both the MUC1 SEA module (SEA, lane 1) andthe mutant uncleaved MUC1 SEA module (SEA-4G, lane 2) were resolved by SDS-PAGE and stained with Coomassie blue stain (panel a). This revealedthe separatea andb subunits of theSEAmodule (lane 1, indicatedbya andb) or thea andb subunits connectedby the4-glycine residue linker (a-4G-b). IdenticalgelswereWesternblotted andprobedwithmAbsBOS10D2,DMB5F3,DMB4F4, orDMB4B4 (panelsb–e, respectively). Theepitope recognizedby theBOS10D2antibodies has been previously reported (26) and is indicated in the bottom panel. Color scheme of the protein segments is the same as in Fig. 1.






that all 5 Ig-g1 mAbs differ one from the other in their bindingcharacteristics.

The epitopes of anti-SEAmodule monoclonal antibodiesDMB5F3, DMB4F4, and DMB4B4 are largelyconformational and involve elements contributed byboth the a and the b subunits

To define the nature of the epitopes bound by the anti-SEAantibodies, they were assessed by Western blotting with aseries of SEA module constructs. Staining of SDS-PAGE gelsof the SEA module (Fig. 4B, a) shows its cleaved constituenta and b chains, as well as protein SEA-4G, a mutant constructconsisting of uncleaved a/b bound by a 4-glycine peptide (6).As expected, probing Western blots with the previously de-scribed anti-b subunit–specific mAb BOS10D2 (ref. 26; linearepitope indicated in bottom schema, Fig. 4B) shows reactivitywith the b subunit but none at the position of the a subunit(Fig. 4B, lane 1). It further shows strong binding to the a-4G-bconstruct (Fig. 4B, lane 2) as well as to small amounts of un-

cleaved a/b protein (ref. 4; Fig. 4B, lane 1). Antibodies DMB5F3and DMB4F4 in contrast react strongly with uncleaved a-4G-bbut are nonreactive with both the a and b subunits (Fig. 5C andD, respectively). Removal of SDS during the blotting procedurelikely allows uncleaved a-4G-b protein to adopt a structuresimilar to that of the SEA module, previously shown to form anunusually stable 3-dimensional structure (6). In contrast, panel Eshows that antibody DMB4B4 reacts with none of the proteins,not witha, not with b subunits in isolation, nor with thea-4G-bmutant isoforms, suggesting that its binding site requires thepresence of both thea and b chains in an epitope resulting fromconformationally determined cleavage.

Affinity of antijunctional antibodiesHaving shown the specificity of the DMB monoclonals to a

MUC1 cell-bound domain, the antibodies were examinedby serial dilutions to evaluate their affinity for intact in situcell-surface MUC1. To do so, the antibodies were reactedwith ZR75 MUC1þ breast cancer cells in flow cytometry

–50

0

50

100

150

200

250

2,5002,0001,5001,0005000

Time

Resp. D

iff.

RU

–30–24–18–12–606

121824303642485460

2,0001,5001,0005000

Time

Resp. D

iff.

A

B

4.84 X 10-101.26 X 10-32.37 X 106

KD (mol/L)Kd (1/s)Ka (1/ms)

5.89 X 10-127.39 X 10-61.25 X 106

DMB5F3

mAb

DMB5F3

DMB4B4

DMB4F4

4 μg/mL

DMB5F3

DMB4B4

DMB4F4

1 μg/mL

DMB5F3

4 μg/mL

1 μg/mL

250 ng/mL

62 ng/mL

ZR75 breast cancer cells

KD (mol/L)Kd (1/s)Ka (1/ms)

DMB5F3

Fab

Figure 5. Titration analysis of DMBmonoclonal anti-MUC1 SEAmodule a/b junction antibodies tohuman breast cancer cellsexpressing MUC1 and affinityanalysis. MUC1þ human ZR75breast cancer cells were reactedwith antibodies DMB4B4,DMB4F4, and DMB5F3 at finalconcentrations of 4 and 1 mg/mL(left and center, respectively), andantibody binding was assessed byflow cytometry. Antibody DMB5F3was shown to have the highestaffinity. Its reactivity with ZR75cells was assessed at decreasingconcentrations (as indicated,right), down to 62 ng/mL,equivalent to a 0.4 nmol/Lconcentration of DMB5F3 (right).SPR binding assays for thedetermination of DMB5F3 full mAband Fab fragments (bottom A andB, respectively) were carried out asdescribed in Materials andMethods.

Pichinuk et al.





(Methods). ZR75 was reacted with the antibodies at an ini-tial concentration of 4 mg/mL. As seen in Fig. 5 (top panel,left), all 3 antibodies, DMB4B4, DMB4F4, and DMB5F3,showed binding at that concentration. At 1 mg/mL bindingof DMB4B4 and DMB4F4 was diminished, whereas DMB5F3clearly showed binding (Fig. 5, middle panel). DMB5F3continued to bind the ZR75 cells to a concentration as lowas 62 ng/mL (Fig. 5, right panel).

Binding affinity of DMB5F3 by surface plasmonresonanceAs DMB5F3 showed the highest binding affinity, we con-

centrated further studies on this antibody. An SPR bindingassay was carried out with serial dilutions of DMB5F3 rangingfrom 7 to 0.4 nmol/L as described in Methods. Over a period of2,250 seconds, the degree of dissociation is almost impercep-tible in each of the 7 curves shown (Fig. 5). These studies wereextended to a partially humanized chimeric version ofDMB5F3(chDMB5F3, see below), and showed that for chDMB5F3 anassociation rate (Ka 1/ms) of 1.25� 106 and a dissociation rate(Kd 1/s) of 7.37� 10�6 were obtained, and an extraordinary lowdissociation constant (KD) of 5.89 � 10�12 mol/L was calcu-lated (panel A). To determine the binding kinetics of the Fabfragment, DMB5F3 Fab was similarly subjected to the SPRbinding assay using Biacore3000 with serial dilutions. Analysisof chDMB5F3 Fab (purified as in Methods) by the SPR bindingassay showed an association rate (Ka 1/ms) of 2.37 � 106, adissociation rate (Kd 1/s) of 1.26 � 10�3, and a dissociationconstant (KD) of 4.84 � 10�10 mol/L (Fig. 5B). Given theincreased avidity associated with a bivalent IgG, the 100-foldKD enhancement seen in the full immunoglobulin molecule ascompared with the Fab is as expected.

Generation of humanized DMB5F3 in Chinese hamsterovary cellsAs an initial demonstration of its clinical application, the

picomolar affinity DMB5F3 antibody was partially human-ized. To generate humanized DMB5F3 mAb, the H andL chains were isolated from SDS-PAGE gels, and theirN-terminal sequences (10–15a.a.) determined. On the basisof that amino acid sequence, nucleotide primers were gen-erated for use in PCR of DMB5F3 cDNA together withprimers generated from known human Ig H and L constantregion sequences. Analysis of the heavy chain amino acidsequence revealed it to be a member of the VH3 gene family,differing in a total of 9 amino acids from germline geneVH36-60/A1/85 (not shown). Of the 9 mutations, 5 arelocated in the CDR regions, whereas 2 are present near theusually highly conserved N terminal of the VH FR1 segment;one of those mutations (from L to V) occurs at a.a. 4 from theFR1 N terminal. The DMB5F3 VL chain is a k light chain,containing 5 a.a. mutations compared with its closest germ-line sequence 23 to 48. Of the 5 mutations, 3 are present inthe CDR1 and CDR3 regions.To show that chimeric DMB5F3 generated by mammalian

cells retains its specific anti-MUC1 activity, recombinantchDMB5F3 was produced in CHO cells. To assess chDMB5F3immunoreactivity, wells were coated with MUC1-Xex protein

and reactedwith dilutions of either recombinant chDMB5F3 ormouse hybridoma DMB5F3 mAb (Fig. 6C and D, respectively).Both antibodies bind the MUC1-Xex protein at similar lowpicomolar concentrations (Fig. 6D). Flow cytometry analysesconfirmed both the specificity and remarkably strong affinityof the recombinant chDMB5F3 antibody to the MUC1 protein.No binding was observed to MUC1-negative parental non-transfected DA3 cells, whereas robust reactivity was seen withmouse transfectants expressing human MUC1-TM (DA3-TM; Fig. 6E and F, respectively). In fact, binding was still clearlyobserved at a concentration of chDMB5F3 as low as 20picoMolar.

Internalization of the DMB5F3 mAbTo assess degree of internalization of the DMB5F3 mAb and

thereby its ability to transport bound drugs into the cell,MUC1-expressing cells were incubated with DyLight649-labeled DMB5F3 either at 4�C or at 37�C and fluorescencemonitored by confocal laser microscopy (Fig. 6G and H,respectively). At 4�C, labeled DMB5F3 was restricted to thecell membrane, indicating binding to the cell surface only.Subsequent incubation at 37�C showed relocalization of thelabeled antibody to within the cell, reflecting highly efficientinternalization to the cell interior.

Linkage of chDMB5F3 generated byCHO toPseudomonasZZ-PE38 toxin and cytotoxicity of the immunoconjugate

Having seen that the DMB5F3 antibody undergoes effi-cient intracellular internalization, we then assessed its abil-ity to ferry cytotoxic moieties into the cell. The ZZ-PE38fusion Pseudomonas toxin is devoid of a cell-binding domain,and to effect its cell killing activity, it must be linked to anagent capable of internalizing it into the cell. ChimericDMB5F3 (chDMB5F3) contains human Fc to which the ZZdomain (derived from protein A) of the ZZ-PE38 toxin bindswell. Parental, nontransfected cells that do not expresshuman MUC1 are unaffected by chDMB5F3-ZZ-PE:38 immu-notoxin (Fig. 6I, diamond symbols), whereas DA3 transfec-tants expressing human MUC1-TM (DA3-TM) are exquisitelysensitive to chDMB5F3:ZZ-PE38 conjugate with cytocidalactivity seen at picoMolar concentrations (Fig. 6I, squaresymbols and see inset). These findings were extended toT47D human breast cancer cells that natively express theMUC1 protein. The chDMB5F3:ZZ-PE38 immunotoxin con-jugate was potently cytocidal to these cells as well, withextensive cell killing observed at as low as picoMolar con-centrations (Fig. 6I).

Efficacy and specificity of cell killing by chDMB5F3:ZZ-PE38 immunotoxin conjugates and comparison ofcytocidal activity with cetuximab (Erbitux) andtrastuzumab (Herceptin)

As seen with DA3 cells transfected with human MUC1 (Fig.6G and H), human adenocarcinoma cells showed that follow-ing incubation at 4�C labeled DMB5F3 was limited to the cellmembrane, indicating binding only to the cell surface (Fig. 7A).Subsequent incubation at 37�C showed almost complete relo-calization of the labeled antibody to within the cell, reflecting






highly efficient internalization into the cell interior (Fig. 7A0).Having shown internalization of chDMB5F3 to MUC1-positivehuman adenocarcinoma cells, we next proceeded to assess thecytotoxicity of immunotoxin formed with chDMB5F3 and tocompare its activity with that of Erbitux and Herceptin immu-notoxins. Fixed cells showed similar cell-membrane reactivitywith both chDMB5F3 and Erbitux (Fig. 7B and B0, respectively),and with Herceptin (data not shown). RecombinantchDMB5F3:ZZ-PE38 reacted with the MUC1þ human pancre-atic cancer cell line Colo357 resulted in cell killing, and serialdilutions of the antibody revealed an IC50 of approximately 16pmol/L (Fig. 7C). However, when chDMB5F3:ZZ-PE38 wasreacted with Colo357 simultaneously with soluble recombi-nant extracellular domain of MUC1-X (MUC1-Xex), competi-tion was seen with a resultant marked reduction in cell killing(Fig. 7C, orange dotted curve). To show the specificity of thatcompetition, solubleMUC1 Xex was added to Erbitux:ZZ-PE38

at an equal concentration.MUC1Xex failed to abrogate Erbituximmunotoxin activity (Fig. 7C, blue dotted curve). In a similarway, serial dilutions of chDMB5F3:ZZ-PE38 reacted with ZR75breast cancer cells showed highly effective killing with an IC50

of 3 pmol/L, whereas addition of MUC1 Xex competitorabolished cell killing (Fig. 7D, compare continuous and dottedorange lines). Whereas reaction of Herceptin:ZZ-PE38 to ZR75cells resulted in cell killing, addition of MUC1 Xex did nothingto abrogate cell death (Fig. 7D, compare continuous and dottedgreen lines). Cell lines such as A431 (epidermoid carcinoma)and N87 (gastric carcinoma) that were found to be only veryslightly positive for MUC1, as assessed by flow cytometry withDMB5F3 mAbs, were not affected by the chDMB5F3:ZZ-PE38immunoconjugate (data not shown).

Taken together, these results showed the specificity of cellkilling mediated by the chDMB5F3:ZZ-PE38 immunotoxin,and furthermore, showed that in certain cancer lines, anti-

1,300 pmol/L

300 pmol/L

80 pmol/L

20 pmol/L

1,300 pmol/LFE

1:10

1:20

1:40

1:80

62 ng/mL

31 ng/mL

15 ng/mL

8 ng/mL

hIgG

standard

CHO recomb.

hDMB5F3

20

10

5

2.5

hDMB5F3

CHO recomb.

25

pmol/L

12.5

6

3

mDMB5F3

hyb. SM

DCBA

pmol/L

40

60

80

100

10 30 50 70 90

Cell

via

bili

ty (

%age)

100

80

60

40

80

700600500400300200100

Antibody concentration (pmol/L)

DA3-PAR DA3-TM T47D

HG

I

4oC 37oC

Figure 6. Activity of recombinantchimeric DMB5F3 (chDMB5F3)mAb synthesized in CHO cells.Recombinant antibody(chDMB5F3, see Materials andMethods) secreted into CHOculture medium was assessed bydot-blotting dilutions (10-fold to80-fold) followed by detection withHRP-conjugated anti-human Fc(B). To assess absolute amounts ofchDMB5F3, a parallel dot blot wascarried out with known amounts ofhuman IgG (A). The activity of thechDMB5F3 recombinant antibodywas compared with that of mousehybridoma DMB5F3 mAb byELISA (C andD, respectively). Flowcytometry analyses were carriedout with the chDMB5F3 antibodyat the indicated concentrationsusing MUC1-negative parentalDA3 cells (DA3-PAR) and mousecell transfectants (DA3-TM)expressing human MUC1-TM (Eand F, respectively). MUC1þ cellswere incubated with DyLight649-labeled DMB5F3 mAb at 4�C andat 37�C (G andH, respectively) andphotographed by confocal lasermicroscopy. Whereas at 4�Cantibody remained at the cellsurface, at 37�C DMB5F3 was allbut completely internalized into thecell. I, cytocidal activity ofchDMB5F3:ZZ-PE38immunotoxin conjugate wastested on DA3-PAR, DA3-TM, andhumanbreast cancer T47Dcells byapplying varying levels ofchDMB5F3 antibody, as indicatedon the x-axis, together withZZ-PE38. Cell viability (y ordinate)was assessed using the alkalinephosphatase assay (see Materialsand Methods).

Pichinuk et al.





MUC1-SEA module DMB5F3 immunotoxin is the most potentof the toxin conjugates tested.

DiscussionDespite its potential to serve as an effective immunother-

apeutic agent for a variety of high-expressing MUC1 malig-nancies, the MUC1 molecule has proved to be an elusiveprey. This is in large measure due to the failure to identifyand target MUC1 moieties that are permanently cell bound(27).Almost all anti-MUC1 antibodies reported to date are

directed against the highly immunogenic polymorphic arrayof 20 to 125 tandem repeats of a 20-amino acid sequencepresent in the MUC1 a chain (for example; refs. 17, 28–30). Asthe a chain is bound noncovalently with the cell-bound bsubunit, it is often shed from the surface of MUC1þ cells andfreely circulates peripherally. It is here that the sheda subunitssequester antitandem repeat array antibodies, thereby pre-cluding their ability to reach MUC1þ tumor cells (15, 16).We previously described mechanisms whereby the cleaved

junction composed of the MUC1 a and b chains is formed (4).Specifically, we analyzed the potential for cleavage of anMUC1-X isoform that contains the same intracellular andmembrane domains as the fullMUC1-TMmolecule (4).Where-as the extracellular domain of MUC1-TM contains the highlyimmunogenic tandem repeat array, the extracellular domain ofMUC1-X is comprised solely of the 120-amino acid SEAmodule

fused to a 30 N-terminal amino acid segment of MUC1,resulting in a less complex structure. Significantly, not onlyis the MUC1-X a/b junctional isoform cleaved, it is cleaved atan identical site as is the full-length MUC1-TM molecule andtherefore results in the same noncovalent interaction of the aand b subunits (4).

All 7 anti-MUC1 mAbs described here are directed againstthe cell-bound MUC1 SEA domain that embraces the MUC1a/b junction. Because the membrane-bound a/b junctioninvolves an intricate structure of a helices and intertwinedb strands (6), antibodies directed against the native structurewould likely recognize epitopes composed of elements con-tributed by both the a and b subunits. As shown (Fig. 4), this isin fact the case: binding by all antibodies requires intactconformation involving elements contributed by both the aand the b subunits. The antibodies not only conform to theexpectation of conformational binding, but they also bind theintact, native glycosylated MUC1 molecule on MUC1þ breastcancer cells (Fig. 5) andMUC1þ pancreatic cancer cells (Fig. 2),with very high affinity—aKD of approximately 6 picoMolar wascalculated for DMB5F3 IgG binding to MUC1 protein. Consid-ering the large number of tandem repeat epitopes in theMUC1a subunit as compared with the unique epitopes present in theSEA domain, the high affinity of the anti-SEA compared withanti-a antibodies is all themore impressive. The DMB5F3 KD is3 to 4 logs higher than the KD values reported for eithercetuximab (31) or trastuzumab (32), both humanized IgGsderived from murine hybridomas, and 3 logs higher than an

0

20

40

60

80

100

120

1,0001001010.10.010.0010

20

40

60

80

100

120

1,0001001010.10.010.001

ZR75 breast cancerColo357 pancreatic cancer

DMB5F3-IT

Erbitux-IT

Erbitux-IT + MUC1-Xex

DMB5F3-IT

Herceptin-ITHerceptin-IT + MUC1-Xex

DMB5F3-IT + MUC1-Xex

IC50

~ 16 pmol/L

IC50

~ 3 pmol/L

Concentration of antibody-IT (pmol/L)Concentration of antibody-IT (pmol/L)

DC

Cell

via

bili

ty (

perc

ent)

Cell

via

bili

ty (

perc

ent)

DMB5F3-IT + MUC1-Xex

A A’ B B’

Figure 7. Specificity and affinity of chDMB5F3 immunotoxin (IT) for cell killing of cancer cells—comparison with Erbitux and Herceptin immunotoxins (IT).Human adenocarcinoma cells were incubated at 4�C (A) followed by incubation at 37�C (A0 ) with DyLight650-labeled DMB5F3 mAb—internalization of theantibody is clearly seen at the 37�C incubation. Unambiguous membrane labeling was seen following incubation of fixed human adenocarcinoma cellswith either chDMB5F3 (B, green label, detection with Alex488-labeled goat anti-human Fc) or Erbitux (B0, red label, detection with R-phycoerythrin-labeledgoat anti-human Fc). Human Colo357 pancreatic and ZR75 breast cancer cells (C and D, respectively) were incubated with chDMB5F3, Erbitux, orHerceptin ZZ-PE38 immunotoxin conjugates, and cell viability (y ordinate) was assessed using the alkaline phosphatase assay. Addition of 50 mg/mL ofsoluble MUC1 Xex to chDMB5F3 immunotoxin abrogated cell killing (dotted orange lines). In contrast, addition of MUC1-Xex competitor did not abrogatecell killing mediated by immuntoxin conjugates formed with either Erbitux (C, compare dotted blue and solid blue lines) or Herceptin (D, compare dottedgreen and solid green lines).






anti-MUC1 tandem repeat array IgG1 isolated from in vitroselection of an Fab phage library.

Naked antibodies against carcinoma cell surface tumorantigens are seldom curative by themselves and are usuallyadministered in combination with chemotherapy. However,very potent cytotoxic agents are often, as stand-alone ther-apeutics, too toxic and should be ideally directed solely tothe tumor antigen expressing cancer cell. This can beachieved via immunoconjugates that will selectively deliverpotent therapeutics to a tumor, thereby reducing systemictoxicity.

As an initial demonstration of ultimate clinical applica-tion, the DMB5F3 anti-MUC1 was partially humanized andproduced in CHO cells as chimeric DMB5F3 (chDMB5F3),and then tested for cytocidal activity as an immunotoxin.Expression of chDMB5F3 in CHO cells showed that itretains the same specificity and picomolar binding affinityas antibody synthesized by the original hybridoma cells. Theresults show that chDMB5F3 not only binds MUC1þ malig-nant cells but that as an immunoconjugate, it also inter-nalizes ZZ:P38 Pseudomonas toxin resulting in very potentcell death (Fig. 6). This is consistent with the previouslyreported cytocidal activity of PE38-immunconjugatesagainst pancreatic cancer cells and against hairy cell leu-kemia (33, 34). In addition, the cytotoxic activity ofchDMB5F3 immunoconjugate was compared with that ofPE38 immunotoxins formed with cetuximab (Erbitux; anti-EGFR; ref. 35) and with trastuzumab (Herceptin, anti-ErbB2or Her2; ref. 36). In human cancer cell lines assessed fortheir sensitivity to the 3 immunotoxin conjugates,chDMB5F3:ZZ-PE38 resulted in cytotoxicity comparable orsuperior to the cetuximab and trastuzumab conjugates inpancreatic and breast cancer cells (Fig. 7).

To assess the antitumor activity of the chDMB5F3:ZZ-PE38immunoconjugate within the intact organism, we have recent-ly initiated a study that uses a nudemousemodel of pancreatictumor growth using MUC1-positive human pancreatic cancercells. To date, our in vivo findings show abrogation oftumor growth corroborating the potent in vitro cytocidalactivity of chDMB5F3:ZZ-PE38 (to be presented in a separatepublication).

The anti-CD30 immunoconjugate, Brentuximab vedotin,has been recently approved for use in CD30þHodgkins disease(37), in which more than a third of patients with refractoryHodgkin lymphoma achieved complete remission and partialremissions were observed in another 40%. This is furtherindication of the effective targeting capabilities of immuno-conjugates and highlights the clinical potential of such anti-body–drug conjugates.

Cetuximab-induced tumor cytotoxicity has not beenfound to directly correlate with the degree of tumor expres-sion of epidermal growth factor receptor (EGFR), but to beprimarily effective in the subset of EGFRþ tumors havingthe KRAS mutation in colorectal adenocarcinoma andsquamous cell carcinoma of the head and neck (38, 39).Furthermore, cetuximab has shown limited activity againstother EGFR-expressing tumors such as breast cancer (40).The antitumor effect of trastuzumab, on the other hand,

does in general correlate with tumor overexpression ofits target protein, ErbB2 (Her2; ref. 36). However ErbB2is overexpressed in only a minority subset of patients,approximately 20% of breast cancer (36) and 22% of gastriccancers (41). Because MUC1 is overexpressed in some 70%to 80% of adenocarcinomas including gastric, pancreatic,colorectal, prostate, and ovarian cancers, as well as on themalignant plasma cell of multiple myeloma (8–13), the highdegree of chDMB5F3:immunotoxin conjugate–inducedcytotoxicity showed here augurs well for use in thosetumors not amenable or nonresponsive to therapy withother mAbs.

Additional investigations, beyond the scope of this article,will be required to see whether the anti-MUC1-SEA domainDMB mAbs reported here possess intrinsic antitumor activ-ity within the organism when administered as naked anti-bodies. Such activity shown for cetuximab and trastuzumabhas been attributed to immunologically based and Fc-depen-dent mechanisms (42), including ADCC (antibody-depen-dent cell-mediated cytotoxicity) and CDC (complement-dependent cytotoxicity). We do not know as yet whetherthis will be the case for all or any of the DMB mAbs reportedhere. Antibodies against the tandem repeat part of MUC1 failto induce complement lysis despite appreciable binding byflow cytometry. This has been attributed to the great dis-tance from the cell surface that complement is activated(43). In this regard, it is pertinent to note that all DMB mAbsbind MUC1-SEA domain epitopes that are located close by tothe cell membrane making CDC a distinct possibility forDMB mAbs.

Tumor stem cells comprise a cell subpopulation capable ofself-regeneration and are thought to result in tumor cellrepopulation following ablation of all detectable malignantdisease by antitumor therapies (44). The present findingsindicate that the side population of MCF7 cells previouslyshown to bear some characteristics of cancer stem cells (22)express MUC1þ. The fact that antibodies DMB4B4 andDMB4F4 bind this side population of MCF7 cells suggests thatthe same conformationally determinedMUC1 epitopes formedby the a and b chains on differentiated tumor cells are alsopresent on these cells.

Targeting conformationally determined sites on the cell-bound a/b MUC1 junction may successfully overcome thedifficulty encountered to date in targeting the MUC1 shed achain. Whether the ultimate optimal clinical application arethe antijunctional antibodies described here or development ofanti-a/b MUC1 junction T-cell vaccines (44–48)—or both—will be answered by further studies.

Disclosure of Potential Conflicts of InterestD.H. Wreschner has ownership interest in Biomodifying, Inc.

Authors' ContributionsConception and design: E. Pichinuk, C. Sympson, D.B. Rubinstein, D.H.WreschnerDevelopment of methodology: E. Pichinuk, I. Benhar, C. Sympson, D.H.WreschnerAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): E. Pichinuk, O. Jacobi, M. Chalik, R. Ziv, A. Karwa, N.I.Smorodinsky, D.H. Wreschner

Pichinuk et al.





Analysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): E. Pichinuk, C. Sympson, D.H. WreschnerWriting, review, and/or revisionof themanuscript:E. Pichinuk, I. Benhar, C.Sympson, D.B. Rubinstein, D.H. WreschnerAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): E. Pichinuk, L. Weiss, R. Ziv, D.B.Rubinstein, D.H. WreschnerStudy supervision: C. Sympson, D.H. Wreschner

AcknowledgmentsThe authors particularly thank Drs. Olivera J. Finn and Dr. Katja Engelmann

from University of Pittsburgh School of Medicine for assistance with the MCF7side population cell experiments.

Grant SupportThis research was funded in part by grants from the Israel Cancer Association

(Project Number 20112024) and Israel Science Foundation (Project Number1167/10).

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicate thisfact.

Received January 12, 2012; revised March 29, 2012; accepted April 5, 2012;published OnlineFirst April 16, 2012.

References1. Baruch A, Hartmann M, Zrihan-Licht S, Greenstein S, Burstein M,

Keydar I, et al. Preferential expression of novel MUC1 tumor antigenisoforms in human epithelial tumors and their tumor-potentiatingfunction. Int J Cancer 1997;71:741–9.

2. Gendler SJ, Lancaster CA, Taylor-Papadimitriou J, Duhig T, Peat N,Burchell J, et al. Molecular cloning and expression of human tumor-associated polymorphic epithelial mucin. J Biol Chem 1990;265:15286–93.

3. Ligtenberg MJ, Vos HL, Gennissen AM, Hilkens J. Episialin, a carci-noma-associated mucin, is generated by a polymorphic gene encod-ing splice variants with alternative amino termini. J Biol Chem1990;265:5573–8.

4. Levitin F, Stern O,WeissM, Gil-Henn C, Ziv R, Prokocimer Z, et al. TheMUC1 SEA module is a self-cleaving domain. J Biol Chem 2005;280:33374–86.

5. Lillehoj EP, Han F, Kim KC. Mutagenesis of a Gly-Ser cleavage site inMUC1 inhibits ectodomain shedding. BiochemBiophys Res Commun2003;307:743–9.

6. Macao B, Johansson DG, Hansson GC, Hard T. Autoproteolysiscoupled to protein folding in the SEA domain of the membrane-boundMUC1 mucin. Nat Struct Mol Biol 2006;13:71–6.

7. LigtenbergMJ, Kruijshaar L, Buijs F, vanMeijerM, Litvinov SV, HilkensJ. Cell-associated episialin is a complex containing two proteinsderived from a common precursor. J Biol Chem 1992;267:6171–7.

8. O'Connor JC, Julian J, Lim SD, Carson DD. MUC1 expression inhuman prostate cancer cell lines and primary tumors. Prostate CancerProstatic Dis 2005;8:36–44.

9. Yonezawa S, Byrd JC, Dahiya R, Ho JJ, Gum JR, Griffiths B, et al.Differential mucin gene expression in human pancreatic and coloncancer cells. Biochem J 1991;276(Pt 3): 599–605.

10. Lapointe J, Li C, Higgins JP, van deRijnM,Bair E,Montgomery K, et al.Gene expression profiling identifies clinically relevant subtypes ofprostate cancer. Proc Natl Acad Sci U S A 2004;101:811–6.

11. Joshi MD, Ahmad R, Yin L, Raina D, Rajabi H, Bubley G, et al. MUC1oncoprotein is a druggable target in human prostate cancer cells. MolCancer Ther 2009;8:3056–65.

12. Gold DV, Karanjawala Z, Modrak DE, Goldenberg DM, Hruban RH.PAM4-reactive MUC1 is a biomarker for early pancreatic adenocar-cinoma. Clin Cancer Res 2007;13:7380–7.

13. Kawano T, Ahmad R, Nogi H, Agata N, Anderson K, Kufe D. MUC1oncoprotein promotes growth andsurvival of humanmultiplemyelomacells. Int J Oncol 2008;33:153–9.

14. Cheever MA, Allison JP, Ferris AS, Finn OJ, Hastings BM, Hecht TT,et al. The prioritization of cancer antigens: a national cancer institutepilot project for the acceleration of translational research. Clin CancerRes 2009;15:5323–37.

15. Prinssen HM,Molthoff CF, Verheijen RH, Broadhead TJ, Kenemans P,Roos JC, et al. Biodistribution of 111In-labelled engineered humanantibody CTM01 (hCTM01) in ovarian cancer patients: influence ofprior administration of unlabelled hCTM01. Cancer Immunol Immun-other 1998;47:39–46.

16. Gillespie AM, Broadhead TJ, Chan SY, Owen J, Farnsworth AP,Sopwith M, et al. Phase I open study of the effects of ascending dosesof the cytotoxic immunoconjugateCMB-401 (hCTMO1-calicheamicin)in patients with epithelial ovarian cancer. Ann Oncol 2000;11:735–41.

17. PegramMD, Borges VF, Ibrahim N, Fuloria J, Shapiro C, Perez S, et al.Phase I dose escalation pharmacokinetic assessment of intravenoushumanized anti-MUC1 antibody AS1402 in patients with advancedbreast cancer. Breast Cancer Res 2009;11:R73.

18. Rubinstein DB, Karmely M, Ziv R, Benhar I, Leitner O, Baron S, et al.MUC1/X protein immunization enhances cDNA immunization in gen-erating anti-MUC1 alpha/beta junction antibodies that target malig-nant cells. Cancer Res 2006;66:11247–53.

19. Mazor Y, Keydar I, Benhar I. Humanization and epitopemapping of theH23 anti-MUC1 monoclonal antibody reveals a dual epitope specific-ity. Mol Immunol 2005;42:55–69.

20. Keydar I, Chou CS, Hareuveni M, Tsarfaty I, Sahar E, Selzer G, et al.Production and characterization of monoclonal antibodies identifyingbreast tumor-associated antigens. Proc Natl Acad Sci U S A1989;86:1362–6.

21. Rubinstein DB, Karmely M, Pichinuk E, Ziv R, Benhar I, Feng N, et al.TheMUC1 oncoprotein as a functional target: immunotoxin binding toalpha/beta junctionmediates cell killing. Int J Cancer 2009;124:46–54.

22. Engelmann K, Shen H, Finn OJ. MCF7 side population cells withcharacteristics of cancer stem/progenitor cells express the tumorantigen MUC1. Cancer Res 2008;68:2419–26.

23. Mazor Y, Barnea I, Keydar I, Benhar I. Antibody internalization studiedusing a novel IgG binding toxin fusion. J Immunol Methods2007;321:41–59.

24. Ali HR, Dawson SJ, Blows FM, Provenzano E, Pharoah PD, Caldas C.Cancer stem cell markers in breast cancer: pathological, clinical andprognostic significance. Breast Cancer Res 2011;13:R118.

25. Kok M, Koornstra RH, Margarido TC, Fles R, Armstrong NJ, Linn SC,et al. Mammosphere-derived gene set predicts outcome in patientswith ER-positive breast cancer. J Pathol 2009;218:316–26.

26. Hartman M, Baruch A, Ron I, Aderet Y, Yoeli M, Sagi-Assif O, et al.MUC1 isoform specific monoclonal antibody 6E6/2 detects preferen-tial expression of the novel MUC1/Y protein in breast and ovariancancer. Int J Cancer 1999;82:256–67.

27. Kufe DW. Mucins in cancer: function, prognosis and therapy. Nat RevCancer 2009;9:874–85.

28. Hamann PR, Hinman LM, Beyer CF, Lindh D, Upeslacis J, Shochat D,et al. A calicheamicin conjugate with a fully humanized anti-MUC1antibody shows potent antitumor effects in breast and ovarian tumorxenografts. Bioconjug Chem 2005;16:354–60.

29. Koumarianou AA, Hudson M, Williams R, Epenetos AA, Stamp GW.Development of a novel bi-specific monoclonal antibody approach fortumour targeting. Br J Cancer 1999;81:431–9.

30. Snijdewint FG, von Mensdorff-Pouilly S, Karuntu-Wanamarta AH,Verstraeten AA, Livingston PO, Hilgers J, et al. Antibody-dependentcell-mediated cytotoxicity can be induced by MUC1 peptide vacci-nation of breast cancer patients. Int J Cancer 2001;93:97–106.

31. Troise F, Cafaro V, Giancola C, D'Alessio G, De Lorenzo C. Differentialbinding of human immunoagents andHerceptin to the ErbB2 receptor.FEBS J 2008;275:4967–79.

32. Thomas SM, Grandis JR. Pharmacokinetic and pharmacodynamicproperties of EGFR inhibitors under clinical investigation. Cancer TreatRev 2004;30:255–68.

33. Du X, Xiang L, Mackall C, Pastan I. Killing of resistant cancer cells withlow Bak by a combination of an antimesothelin immunotoxin and a






TRAIL Receptor 2 agonist antibody. Clin Cancer Res 2011;17:5926–34.

34. Kreitman RJ, Pastan I. Antibody fusion proteins: anti-CD22 recombi-nant immunotoxin moxetumomab pasudotox. Clin Cancer Res2011;17:6398–405.

35. Vincenzi B, Santini D, Tonini G. New issues on cetuximab mechanismof action in epidermal growth factor receptor-negative colorectalcancer: the role of vascular endothelial growth factor. J Clin Oncol2006;24:1957; author reply 1957–1958.

36. Hudis CA. Trastuzumab–mechanism of action and use in clinicalpractice. N Engl J Med 2007;357:39–51.

37. Gualberto A. Brentuximab Vedotin (SGN-35), an antibody-drug con-jugate for the treatment of CD30-positive malignancies. Expert OpinInvestig Drugs 2012;21:205–16.

38. Lievre A, Bachet JB, Boige V, Cayre A, Le Corre D, Buc E, et al. KRASmutations as an independent prognostic factor in patients withadvanced colorectal cancer treated with cetuximab. J Clin Oncol2008;26:374–9.

39. Di Fiore F, Blanchard F, Charbonnier F, Le Pessot F, Lamy A, GalaisMP, et al. Clinical relevance of KRAS mutation detection in metastaticcolorectal cancer treated by Cetuximab plus chemotherapy. Br JCancer 2007;96:1166–9.

40. Modi S, D'Andrea G, Norton L, Yao TJ, Caravelli J, Rosen PP, et al. Aphase I study of cetuximab/paclitaxel in patients with advanced-stagebreast cancer. Clin Breast Cancer 2006;7:270–7.

41. Okines AF, Cunningham D. Trastuzumab in gastric cancer. Eur JCancer 2010;46:1949–59.

42. Clynes R, Takechi Y, Moroi Y, Houghton A, Ravetch JV. Fc receptorsare required in passive and active immunity to melanoma. Proc NatlAcad Sci U S A 1998;95:652–6.

43. Ragupathi G, Liu NX, Musselli C, Powell S, Lloyd K, Livingston PO.Antibodies against tumor cell glycolipids and proteins, but not mucins,mediate complement-dependent cytotoxicity. J Immunol 2005;174:5706–12.

44. Gilewski T, Adluri S, Ragupathi G, Zhang S, Yao TJ, Panageas K, et al.Vaccination of high-risk breast cancer patients with mucin-1 (MUC1)keyhole limpet hemocyanin conjugate plus QS-21. Clin Cancer Res2000;6:1693–701.

45. Ding C, Wang L, Marroquin J, Yan J. Targeting of antigens to B cellsaugments antigen-specific T-cell responses and breaks immune tol-erance to tumor-associated antigenMUC1. Blood 2008;112:2817–25.

46. Ramanathan RK, Lee KM, McKolanis J, Hitbold E, Schraut W, MoserAJ, et al. Phase I studyof aMUC1vaccine composedof different dosesof MUC1 peptide with SB-AS2 adjuvant in resected and locallyadvanced pancreatic cancer. Cancer Immunol Immunother 2005;54:254–64.

47. Lepisto AJ, Moser AJ, Zeh H, Lee K, Bartlett D, McKolanis JR, et al. Aphase I/II study of a MUC1 peptide pulsed autologous dendritic cellvaccine as adjuvant therapy in patients with resected pancreatic andbiliary tumors. Cancer Ther 2008;6:955–64.

48. Karanikas V, Hwang LA, Pearson J, Ong CS, Apostolopoulos V,Vaughan H, et al. Antibody and T cell responses of patients withadenocarcinoma immunized with mannan-MUC1 fusion protein. JClin Invest 1997;100:2783–92.

Pichinuk et al.





2012;72:3324-3336. Published OnlineFirst April 16, 2012.Cancer Res Edward Pichinuk, Itai Benhar, Oded Jacobi, et al. CellsAntibody Targeting of Cell-Bound MUC1 SEA Domain Kills Tumor

Updated version

10.1158/0008-5472.CAN-12-0067doi:

Access the most recent version of this article at:

Cited articles

http://cancerres.aacrjournals.org/content/72/13/3324.full#ref-list-1

This article cites 48 articles, 20 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/72/13/3324.full#related-urls

This article has been cited by 1 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://cancerres.aacrjournals.org/content/72/13/3324To 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-12-0067

http://cancerres.aacrjournals.org/content/72/13/3324.full#ref-list-1

http://cancerres.aacrjournals.org/content/72/13/3324.full#related-urls

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

mailto:[email protected]

http://cancerres.aacrjournals.org/content/72/13/3324


antibody targeting of cell-bound muc1 sea domain kills ... · antibodies was done with horseradish...

Documents