Int. J. Cancer: 82, 525–531 (1999)
`r 1999 Wiley-Liss, Inc.
`
`Publication of the International Union Against Cancer
`Publication de l’Union Internationale Contre le Cancer
`
`RELATIVE CYTOTOXIC ACTIVITY OF IMMUNOTOXINS REACTIVE WITH
`DIFFERENT EPITOPES ON THE EXTRACELLULAR DOMAIN OF THE c-erbB-2
`(HER-2/neu) GENE PRODUCT p185
`Cinda M. BOYER1, Lajos PUSZTAI2, Jon R. WIENER2, Feng Ji XU2, G. Scott DEAN1, Blanche S. BAST1, Kathy C. O’BRIANT1,
`Marilee GREENWALD1, Karen A. DESOMBRE1 and Robert C. BAST, JR.2*
`1Department of Medicine, Duke University Medical Center, Durham, NC, USA
`2Division of Medicine, University of Texas M.D. Anderson Cancer Center, Houston, TX, USA
`
`Different epitopes on the extracellular domain of the
`HER-2 receptor can serve as distinct targets for immunotox-
`ins. To determine the optimal epitope target for immuno-
`toxin therapy, 7 anti-HER-2 ricin A chain murine monoclonal
`immunotoxins, each reactive with different epitopes of HER-2
`receptor, were tested for cytotoxic activity. The immunotox-
`ins produced 1.2–4.6 logs of cytotoxicity in limiting dilution
`clonogenic assays with 2 breast cancer cell lines that overex-
`pressed HER-2. Cytotoxicity did not correlate with immuno-
`globulin isotype, binding affinity, relative position of epitopes
`or internalization of the anti-HER-2 immunotoxins. Interest-
`ingly, the most and least effective immunotoxins bound to
`epitopes in very close proximity. Competitive binding assays
`with unconjugated antibodies have previously indicated that
`our antibodies recognized epitopes that are arranged in a
`linear array. To orient this relative epitope map, deletions
`were prepared in the HER-2/neu gene and these mutant
`constructs were expressed in NIH3T3 cells. Epitope expres-
`sion was determined by antibody binding and radioimmunoas-
`say. Epitopes targeted by the PB3, 454C11 and NB3 antibod-
`ies are localized N-terminal to the epitopes recognized by
`ID5, BD5, 741F8 and 520C9 antibodies. The 2 non-conforma-
`tional epitopes PB3 and NB3 were localized to regions
`corresponding to amino acides 78–242 of the p185HER-2 pro-
`tein. Int. J. Cancer 82:525–531, 1999.
`r 1999 Wiley-Liss, Inc.
`
`The protein product of the HER-2/c-erbB-2 oncogene (p185), a
`tyrosine kinase receptor, is an appealing target for serotherapy.
`Approximately one third of breast and ovarian cancers overexpress
`p185 and overexpression is associated with poor prognosis in these
`patients, whereas normal tissues express relatively low levels of
`this receptor (Slamon et al., 1989; Berchuck et al., 1990).
`Antibodies against HER-2 either alone or in combination with
`chemotherapy have shown promising results in clinical studies
`(Cobleigh et al., 1998; Slamon et al., 1998). Identification and
`functional analysis of epitopes recognized by distinct anti-HER-2
`antibodies may be important for optimizing this therapeutic
`strategy. Characterization of immunogenic regions of the HER-2
`protein is also important to design strategies to enhance immune
`response against
`this tumor-associated antigen, and functional
`analysis of epitopes may reveal important relationship between
`HER-2 protein structure and function.
`Various anti-HER-2 monoclonal antibodies (MAbs) inhibit cell
`growth in vitro with varying efficiency (Drebin et al., 1988;
`Hudziak et al., 1989; Tagliabue et al., 1991). The functional
`activity of an anti-HER-2 antibody is more related to the epitope
`that it recognizes than to its antigen binding affinity or ability to
`block ligand binding (Xu et al., 1993). Antibodies that bind near the
`transmembrane region of the HER-2 extracellular domain may
`have more potent antiproliferative activity than those binding the
`N-terminus (Lewis-Phillips et al., 1998). On the other hand,
`antibodies that bind close to the N-terminus exert more cytotoxic
`effects than antibodies that bind near the transmembrane region
`(Lewis-Phillips et al., 1998). Immunotoxins directed against dis-
`tinct p185 epitopes may also exert different levels of cytotoxicity
`(Rodriguez et al., 1993; Tecce et al., 1993). Such differential
`cytotoxicity has been demonstrated for immunotoxins reactive with
`the CD2 (Press et al., 1988) and IgD molecules (May et al., 1990).
`
`Conjugates that recognize epitopes on the C-terminus of the CD2
`molecule are more effective than those that bind the N-terminus.
`For IgD, anti-Fc immunotoxins have greater cytotoxicity than
`anti-Fd immunotoxins. For both molecules, epitopes more proxi-
`mal to the cell membrane provide the most effective targets,
`possibly permitting more efficient translocation of toxin moieties.
`Overexpression of p185 on breast and ovarian cancer cell lines
`was required for effective kill with TA1-RTA (Rodriguez et al.,
`1993). As only the TA1 immunotoxin was evaluated, it was not
`clear whether antibodies that bound to different p185 epitopes
`would provide more potent immunotoxins. In this study, we have
`characterized the cytotoxic activity of 7 distinct anti-HER-2
`immunotoxins directed against different epitopes on the extracellu-
`lar domain of p185HER-2. We attempted to determine whether
`cytotoxicity is related to antigen binding affinity, epitope location
`or efficiency of internalization of the immunotoxin. By expressing
`deletion mutants of the extracellular domain of HER-2 protein in
`NIH3T3 cells, we investigated if amino acid sequence information
`could be linked to an epitope map previously generated by
`competitive binding assays.
`
`MATERIAL AND METHODS
`
`Cell lines
`SKBr3 and BT474 human breast cancer cell lines were main-
`tained in RPMI-1640 medium supplemented with 15% heat-
`inactivated fetal bovine serum (FBS) and 2 mM L-glutamine, 100
`units/ml penicillin and 100 µg/ml streptomycin. The NIH3T3
`murine fibroblast cell line and the 17313 subclone of this cell line
`(McKenzie et al., 1989) were kindly provided by Dr. S. McKenzie
`(Cambridge, MA). The 17313 clone was produced by transfection
`of the full-length human neu gene into NIH3T3 cells. Both murine
`cell lines were cultured in Dulbecco’s modified Eagle’s medium
`(DMEM) supplemented with 10% FBS and 2 mM L-glutamine.
`Medium for 17313 was additionally supplemented with 400 µg/ml
`G418 (GIBCO BRL, Grand Island, NY). The additional transfec-
`tants of NIH3T3 described in this report were maintained in
`medium identical to that described for the 17313 cell line. The
`LTR-1/DNerbB2 transfectant and its parental NIH3T3 line were
`kindly provided by Dr. S. Aaronson (National Cancer Institute,
`Bethesda, MD). The LTR-1/DNerbB2 transfectant contains a
`construct of c-erbB-2 that lacks most of the extracellular domain
`(Di Fiore et al., 1987). The LTR-1/DNerbB2 transfectant and its
`parental NIH3T3 cells were cultured in DMEM supplemented with
`10% FBS and 2 mM L-glutamine.
`
`Grant sponsor: National Cancer Institute, Department of Health and
`Human Services; Grant number: CA 39930.
`
`*Correspondence to: Division of Medicine, Box 92, University of Texas
`M.D. Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, TX
`77030, USA. E-mail: rbast@notes.mdacc.tmc.edu
`
`Received 2 July 1998; Revised 26 February 1999
`
`IMMUNOGEN 2038, pg. 1
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`526
`
`BOYER ET AL.
`
`MAbs
`Murine MAbs, that react with the extracellular domain of p185
`were kindly provided by Dr. S. McKenzie (TA1, BD5, ID5, NB3,
`RC1, RC6, PB3 and OD3) ABT-Oncogene Science, Cambridge,
`MA and by Dr. D. Ring (520C9, 741F8 and 454C11) (Ring et al.,
`1991), Emeryville, CA. The 225 MAb recognizes the extracellular
`domain of the epidermal growth factor receptor and was kindly
`provided by Dr. J. Mendelsohn (Masui et al., 1989). All the murine
`MAbs were of the IgG1 isotype except PB3 (IgG2a), 454C11
`(IgG2a), and OD3 (IgM). MOPC21 (IgG1), H16-L10-4R5 (IgG2a)
`and NS.4.1 (IgM) were obtained from the ATCC (Rockerville, MD)
`and used as isotype-matched controls that did not bind to p185.
`Purified MAbs were prepared from hybridoma-induced ascites
`fluid as described previously (Xu et al., 1993). Competitive binding
`assays reported previously suggest that these different antibodies
`recognize a linear array of epitopes on the extracellular domain of
`p185 (Xu et al., 1993).
`
`Immunotoxin preparation
`Immunotoxins were prepared by conjugation with recombi-
`nantly derived RTA kindly provided by Dr. L. Houston (Emery-
`ville, CA) using 2-iminothiolane (Wawrzynaczak and Thorpe,
`1987). The ratio of toxin/antibody was 3 to 5. Immunotoxin
`preparations were characterized by SDS-PAGE. Purified immuno-
`toxins contained no free antibody or ricin A chain by gel analysis.
`Activity of RTA and antibody-RTA conjugates was determined by
`inhibition of translation in a rabbit reticulocyte lysate in vitro
`translation system (Promega, Madison, WI).
`
`Clonogenic assay
`To study the effect of immunotoxins on growth of breast cancer
`cell lines, an in vitro limiting dilution clonogenic assay was used.
`SKBr3 or BT474 breast cancer cells (106) were incubated at 37°C
`for 3 hr in an atmosphere of 5% CO2 and humidified air on a rotator
`with different concentrations of the immunotoxin conjugates in
`growth medium. The cells were washed twice with growth medium
`and serially diluted 5-fold. A 100 µl portion of each dilution was
`plated in each of the 6 wells within a 96-well flat-bottomed
`microtiter plate. An additional 100 µl aliquot of tissue culture
`medium was added to each well. The cells were incubated for 14
`days at 37°C. Clonogenic growth was determined using an inverted
`
`phase microscope, scoring the number of wells with at least 1
`tumor colony that contained at least 10 cells. Estimates of the
`surviving clonogenic units were calculated according to a modifica-
`tion of the method of Spearman and Karber. Clonogenic elimina-
`tion was calculated by subtracting the log clonogenic units of the
`medium-treated cultures from the log clonogenic units of the
`treated cell culture for each individual cell line.
`
`Deletion constructs of c-erbB-2
`Deletion constructs of the c-erbB-2 gene were prepared in a
`shuttle vector, excised and re-ligated into an expression vector for
`transfection into NIH3T3 cells (Fig. 1). A pUC19-plasmid shuttle
`vector (pUC-HER) was prepared which contained the full-length
`HER-2/c-erbB-2 cDNA in a unique Xho1 restriction site kindly
`provided by Dr. S. Aaronson. The shuttle vector was restricted with
`Msc1 to remove a fragment of 630 bp from the intracellular domain
`of the c-erbB-2 cDNA. This excision was required to remove
`sequences containing restriction sites from the cDNA which would
`have interfered with construction of the deletions in the extracellu-
`lar domain. This deletion does not interfere with the protein reading
`frame, with expression of protein in transfected cells or with
`conformation of the extracellular domain of the expressed c-erbB-2
`protein. Mutations were then generated within the c-erbB-2 cDNA
`as described below, followed by excision with Xho1, and insertion
`by ligation into the appropriate expression vector.
`Construct MMTneo Mro1. The pUC-HER/Msc vector was
`restricted with Mro1 at cDNA position 2219. The 38 overhangs
`were filled in with Klenow and the blunt ends re-ligated. This
`construct created an out-of-frame sequence beginning at amino
`acid 684 and ending with a termination codon at amino acid
`position 700.
`Construct PMx1112 Eco 2-20. The pUC-HER/Msc vector was
`restricted with EcoR5 and Nae1, at positions 408 and 900,
`respectively. The vector containing the remaining c-erbB-2 se-
`quences was then re-ligated. Since both enzymes cleaved between
`codons, the reading frame was maintained with the deletion of 492
`bp, or 164 amino acids.
`Construct MMTneo Hinc 6-17. The pUC-HER/Msc vector was
`restricted with Nae1 and Hinc2 at positions 900 and 1764,
`respectively, removing a fragment of 864 bp. The vector containing
`
`FIGURE 1 – Deletion constructs of the c-erbB-2 gene. Deletions were prepared as described in Material and Methods.
`
`IMMUNOGEN 2038, pg. 2
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`CYTOTOXICITY OF ANTI-p185c-erbB-2 IMMUNOTOXINS
`
`527
`
`the remaining c-erbB-2 sequences was re-ligated, resulting in the
`maintenance of the reading frame and the loss of 288 amino acids.
`Construct MMTneo Bgl 5-33. The pUC-HER/Msc vector was
`restricted with Bgl2, which cleaves at cDNA positions 615 and
`1280, removing a fragment of 666 bp. The vector containing the
`remaining c-erbB-2 sequences was re-ligated, resulting in the
`maintenance of the reading frame and the loss of 222 amino acids.
`The constructs were excised from the shuttle vector with Xho1,
`ligated into the MMTneo or PMx1112 expression vectors, and
`determined to have the correct orientation by use of the unique
`Kpn1 site at position 3250. Verification of the deletions in the
`different constructs was performed by sequence analysis.
`
`Transfection
`Deletion constructs of c-erbB-2 were transfected into NIH3T3
`cells using calcium phosphate. In the case of the PMx1112
`expression vector, NIH3T3 cells were cotransfected with the
`PMx1112 vector and the SV2neo vector containing the neomycin
`resistance gene (Dr. R. Kaufman, Durham, NC) at a 5 to 1ratio of
`PMX112 DNA to SV2neo DNA. Transfectants were selected in
`medium containing 400 µg/ml neomycin sulfate (G418; GIBCO
`BRL).
`
`Immunohistochemical staining
`Transfectants containing constructs with deletions in the extracel-
`lular domain of p185 were screened initially for the presence of the
`intracellular domain with the 135C6EE and 145WWA-1 MAbs
`kindly provided by Dr. C. Nolan (Chicago, IL) which were
`produced against a peptide containing amino acid residues 1056
`through 1072 of the p185 sequence. Transfectants were grown in
`Lab-Tek chamber slides (Nunc, Naperville, IL), and monolayers
`rinsed with PBS, air dried and fixed in cold acetone. Binding of
`135C6EE, 145WWA-1 and MOPC21 (non-specific control) to the
`different transfectants was assessed with the Vectastain ABC Kit
`(Vector, Burlingame, CA). MAbs were diluted to 5 µg/ml for
`immunohistochemical staining. Slides were incubated with diluted
`antibody for 1 hr at room temperature and washed with PBS. Slides
`were then incubated with biotinylated anti-mouse IgG (Vectastain
`ABC kit), washed with PBS and incubated with ABC regent
`(Vectastain ABC kit) according to the manufacturer’s instructions.
`Slides were developed with the enzymatic substrate diaminobenzi-
`dine, washed in tap water and counterstained with methyl green.
`Transfectants in which expression of the construct was documented
`by positive staining with 135C6EE and 145WWA-1 were further
`screened for binding of the different antibodies to epitopes on the
`extracellular domain of p185 by indirect live cell radioimmunoas-
`say.
`
`Indirect live cell radioimmunoassay
`The indirect
`live cell radioimmunoassay was performed as
`described previously (Boyer et al., 1989). Briefly, 1–2 3 104 cells
`were plated in flat-bottomed RemovaCell microtiter plates (Dyna-
`tech, Alexandria, VA) and incubated overnight at 37°C in an
`atmosphere of 5% CO2 and 95% humidified air. Assay wells were
`blocked with RPMI-1640 medium supplemented with 10% BSA
`and 0.08% sodium azide for 1 hr at 37°C. Blocking medium was
`removed and 50 µl MAb diluted to 5 µg/ml in assay medium
`(RPMI-1640 medium supplemented with 1% BSA and 0.08%
`sodium azide) was added to quadruplicate wells. Plates were
`incubated for 1 hr at 37°C. Following 3 washes with assay medium,
`125I-labeled sheep antimouse F(ab8)2 fragments (105 cpm) (NEN,
`Boston, MA) were added in 50 µl of assay medium and plates were
`incubated for 1 hr at 37°C. Plates were washed 3 times with assay
`medium and RemovaCell wells separated and counted for 1 min in
`a gamma counter. Binding ratios were calculated for each antibody
`by dividing the mean cpm with antibody divided by the mean cpm
`for assay medium. The average binding ratio obtained in multiple
`experiments for the non-specific isotype control antibody was
`compared to the average binding ratio for each anti-p185 antibody
`by the Student’s t-test. An epitope was considered retained on a
`
`construct if the average binding ratio of the anti-p185 antibody was
`statistically different ( p , 0.05) from the average binding ratio of
`the control and the amount of binding was at least 2-fold above the
`binding of the control antibody.
`
`Radioiodination of MAbs and immunoconjugates
`MAbs were labeled with Na125I using the iodogen method. In
`brief, 50 µl of phosphate buffer (0.5 M, pH 7.4) were added to a
`15 3 75 mm borosilicate tube coated with 10 µg of iodogen
`(Pierce, Rockford, IL). MAb (50 µg) was added in a volume of 95
`µl PBS (50 mM phosphate buffer, 0.15 M NaCl). Radioiodination
`was initiated by the addition of 0.5 mCi of Na125I (5 µl) and the
`mixture was incubated for 30 min on ice. The protein-bound iodine
`was separated from free 125I by gel filtration on a PD-10 column
`(Pharmacia, Pleasant Hill, CA) that had been equilibrated with
`PBS. A sample of 3 µl from each fraction was counted in a gamma
`counter to measure protein-bound radioactivity. Iodination effi-
`ciency was calculated using the following formula:
`
`Iodination efficiency 5
`
`protein bound cpm
`total cpm
`
`3 100%.
`
`Direct live cell radioimmunoassay and Scatchard analysis
`Cells were plated as if indirect live cell radioimmunoassay were
`to be performed. After overnight incubation, monolayers were
`washed with RPMI-1640 medium supplemented with 1% FBS and
`0.1% sodium azide. Different amounts of 125I-labeled MAb or
`immunotoxin were added in 50 µl aliquots to triplicate cell
`monolayers. Non-specific binding was determined by adding
`different amounts of 125I-labeled MAbs or immunotoxins to empty
`wells. After incubation on ice for 4 hr, unbound antibodies were
`removed by washing the wells 4 times with ice-cold assay medium
`supplemented with 5% FBS and 0.1% sodium azide. Individual
`wells were then detached, and radioactivity was determined in a
`gamma counter. The EBDA program was used to calculate the
`number of binding sites (McPherson, 1985):
`
`Binding sites
`
`5
`
`maximum binding (mol) 3 6.23 3 1023 3 volume (L)
`cell number
`
`.
`
`Internalization of antibodies and immunotoxins
`125I-labeled antibody or immunotoxin (0.25 µg/ml) was added to
`aliquots of 106 SKBr3 cells. After incubation for 1 hr at 4°C to
`allow binding, cells were washed 3 times with RPMI-1640 medium
`with 1% BSA and then either counted immediately to determine the
`total amount of antibody or immunotoxin bound or incubated at
`4°C or 37°C for 1 hr to permit internalization of immunoglobulin.
`To remove antibody still bound to the cell surface, 2.5 mg/ml of
`proteinase K (Sigma, St. Louis, MO) was added to the cells for 1 hr
`at 37°C. The cells were washed 3 times in RPMI-1640 medium
`supplemented with 1% BSA and 0.1% sodium azide. Radioactivity
`associated with cell pellets was counted in a gamma counter. The
`amount of antibody internalized was determined by subtracting the
`cpm obtained after incubation at 4°C followed by protease
`stripping from the cpm obtained after incubation at 37°C for the
`same time interval, followed by protease stripping. The percentage
`internalized was then calculated by dividing the cpm of the
`antibody internalized by the total cpm initially bound to the cell
`surface.
`
`Western blot
`SKBr3 cells (20 3 106) were removed from a tissue culture flask
`by scraping into cold PBS, pelleted at 1,000g for 10 min at 4°C and
`incubated in 0.5 ml of lysis buffer (150 mM NaCl, 5 mM EDTA, 50
`mM Tris, pH 8, containing 4 mM benzamidine, 10 µg/ml leupeptin,
`10 µg/ml pepstatin A, and 1 mM phenylmethylsulfonyl fluoride,
`and 1% Nonidet P-40) on ice for 30 min. The lysate was
`centrifuged at 1,000g for 10 min at 4°C and the supernatant was
`
`IMMUNOGEN 2038, pg. 3
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`528
`
`BOYER ET AL.
`
`FIGURE 2 – Cytotoxic activity of RTA immunotoxins reactive with different epitopes of p185c-erbB-2 and the epidermal growth factor receptor in
`SKBr3 breast cancer cells. Data are expressed as mean log kill for 2–4 experiments at a given concentration for each immunotoxin.
`
`collected. Proteins in the SKBr3 lysate were separated by SDS-
`PAGE in 7.5% polyacrylamide gels under reducing conditions
`(PhastSystem; Pharmacia LKB, Piscataway, NJ). Separated pro-
`teins were electrophoretically transferred to nitrocellulose using the
`PhastTransfer semidry protein transfer kit (Pharmacia). Remaining
`protein binding sites on the nitrocellulose membranes were blocked
`by incubation in Blotto (3% non-fat dry milk powder, 2% normal
`goat serum and 0.1% Tween-20 in PBS) for 1 hr at room
`temperature. Membranes were then incubated for 3 hr with
`anti-p185 antibody diluted to 20 µg/ml in Blotto at room tempera-
`ture with gentle agitation. Membranes were washed 3 times in
`wash buffer (20 mM Tris, 1 M NaCl, 0.05% Tween-20, pH 7.6) for
`10 min with gentle agitation followed by incubation with phospha-
`tase-labeled goat anti-mouse IgG 1 IgA 1 IgM diluted 1:500 in
`Blotto (Kirkegaard and Perry, Gaithersburg, MD) for 1 hr at room
`temperature with gentle agitation. Membranes were washed 3 times
`as described above and developed with the BCIP/NBT phosphatase
`substrate system.
`
`Statistical analysis
`Limiting dilution analysis was performed using a Spearman
`estimate. The mean of the dose-response function for each treat-
`ment was estimated by
`
`TABLE I – CYTOTOXIC ACTIVITY (LOG KILL, MEAN 6 SE) OF RTA
`IMMUNOTOXINS REACTIVE WITH DIFFERENT EPITOPES OF p185c-erbB-2 AND
`THE EPIDERMAL GROWTH FACTOR RECEPTOR IN SKBr3 BREAST
`CANCER CELLS
`
`Cell lines
`
`Cytotoxic activity
`
`5 µg/ml
`
`2.5 µg/ml
`
`1 µg/ml
`
`0.5 µg/ml
`
`TA1-RTA 4.12 6 0.77 3.61 6 0.90 2.82 6 0.76 1.89 6 0.53
`BD5-RTA 2.61 6 0.05 1.75 6 0.47 0.64 6 0.06 0.29 6 0.06
`RC6-RTA 2.51 6 1.27 1.28 6 0.82 0.64 6 0.06 0.17 6 0.06
`PB3-RTA 2.27 6 0.41 2.33 6 0.24 1.81 6 0.41 1.11 6 0.30
`520C9-RTA 3.30 6 1.14 2.91 6 0.51 1.98 6 0.66 1.28 6 0.50
`3.15 6 0.69 2.50 6 0.29 1.28 6 0.35 1.34 6 0.18
`ID5-RTA
`741F8-RTA 2.74 6 0.53 2.21 6 0.58 1.46 6 0.65 1.16 6 0.35
`0.62 6 0.47 0.62 6 0.47 0.73 6 0.40 0.36 6 0.08
`225-RTA
`
`detection (at p 5 0.05, 2-sided) of a 10-fold difference in clono-
`genic units between 2 treatment groups with probability 0.90.
`Corrections for multiple comparisons were made when appropri-
`ate.
`
`RESULTS
`Differential cytotoxicity of immunotoxins reactive with distinct
`p185 epitopes
`To compare the activity of immunotoxins reactive with different
`p185 epitopes, ricin A chain (RTA) conjugates were prepared from
`7 antibodies, each recognizing different epitopes on p185 (Xu et
`al., 1993), and 1 antibody, 225, was directed against an epitope on
`the extracellular domain of the epidermal growth factor receptor.
`Cytotoxicity was measured in limiting dilution clonogenic assays.
`Both SKBr3 and BT474 breast cancer cells overexpress p185 and
`also express low levels of the epidermal growth factor receptor. The
`cytotoxic effect of the different immunotoxins was similar on
`BT474 and SKBr3 cells. TA1-RTA was the most effective immuno-
`toxin, whereas BD5-RTA and RC6-RTA were the least effective
`Anti-epidermal growth factor receptor immunotoxin 225-RTA also
`had low cytotoxic activity. Differences were particularly marked at
`immunotoxin concentrations of 0.5 and 1 µg/ml. Unconjugated
`anti-p185 antibodies are ineffective in limiting dilution clonogenic
`assays (data not shown). Both BD5 and ID5 antibodies react with
`
`ri
`
`oi
`
`k
`
`d n
`
`2
`
`d 2
`
`m 5 X0 1
`
`50
`where X0 5 ln initial dose 5 ln 1021; d5 ln dilution factor 5 ln 5;
`n 5 number of wells at each dilution 5 6; k 5 number of 5-fold
`dilutions; and r1 5 the number of wells with observed growth at the
`ith dilution. The estimated number of clonogenic units per ml was
`then calculated as 0 5 exp (20.57722 2 m). The initial dose (1021
`ml) and the number of dilutions (k 5 9) were chosen to have a high
`probability (.0.99) that r0 5 n and rk 5 0. Tests of significance to
`compare 2 treatments were based on the asymptotic normality of
`the estimates of the means (m), each with variance closely
`approximated by d ln2/n. A ‘‘Z’’ statistic was used to estimate the
`level of significance. A formal comparison was performed in which
`the means and variances were obtained from the Spearman
`estimator; 5-fold dilutions with 6 wells/dilution permitted the
`
`IMMUNOGEN 2038, pg. 4
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`TABLE II – BINDING CHARACTERISTICS OF ANTI-p185 ANTIBODIES BEFORE
`AND AFTER CONJUGATION TO RICIN A CHAIN
`
`TABLE IV – INTERNALIZATION OF 125I-LABELED ANTI-p185 ANTIBODIES
`AND IMMUNOTOXINS1
`
`CYTOTOXICITY OF ANTI-p185c-erbB-2 IMMUNOTOXINS
`
`529
`
`Antibody/immunotoxin
`
`Relative affinity1 (M 2 1)
`
`Binding sites/cell
`(3105)
`
`TAI
`TAI-RTA
`BD5
`BD5-RTA
`RC6
`RC6-RTA
`1Mean 6 SD.
`
`(2.56 6 0.68) 3 1029
`(2.93 6 1.94) 3 1029
`(1.09 6 0.18) 3 1028
`(1.23 6 0.16) 3 1028
`(8.15 6 0.64) 3 1029
`(5.95 6 1.40) 3 1028
`
`8.0
`11.0
`14.8
`41.9
`5.8
`11.0
`
`TABLE III – INHIBITION OF PROTEIN SYNTHESIS IN AN IN VITRO
`TRANSLATION SYSTEM BY DIFFERENT RICIN A CHAIN IMMUNOTOXINS
`AND RICIN A CHAIN ALONE (RTA)1
`
`RTA or
`antibody-RTA 5.6 3 10210 M 5.6 3 10211 M 5.6 3 10212 M 5.6 3 10213 M
`
`RTA
`TAI-RTA
`RC6-RTA
`BD5-RTA
`PB3-RTA
`520C9-RTA
`741F8-RTA
`225-RTA
`
`3
`4
`5
`5
`5
`2
`3
`7
`
`3
`4
`5
`5
`3
`3
`3
`5
`
`16
`14
`35
`16
`5
`5
`6
`11
`
`55
`51
`55
`40
`18
`16
`21
`32
`
`1Percent inhibition of 35S-methionine incorporation into proteins in
`rabbit reticulocyte lysate in vitro translation system. Inhibition is
`expressed as percentage of control (treatment with corresponding
`unconjugated antibody alone) measured at 4 different immunotoxin
`and RTA concentrations).
`
`the extracellular domain of p185 at or near the ligand binding site,
`since these antibodies competitively inhibit
`ligand binding to
`HER-2 receptor and paradoxically also induce growth simulation
`in soft agar (Xu et al., 1993).
`
`Cytotoxicity of immunotoxins is not related to isotype, different
`p185 binding affınity or differential ricin activity
`Six of the 7 antibodies used in these studies are of the IgG1
`isotype and 1, PB3, is of the IgG2a isotype, indicating that cytotoxic
`activity in vitro is not related to antibody isotype. Similarly, the
`relative binding affinity of the different anti-p185 antibodies did not
`correlate with the cytotoxicity of the corresponding immunotoxin.
`Relative binding affinities of 11 different anti-p185 antibodies
`including the 7 utilized in our study have been determined on 2 cell
`lines with differing levels of p185 and reported previously (Xu et
`al., 1993). TA1 demonstrated the highest relative binding affinity
`and the relative binding affinities for the remaining 6 antibodies
`were similar.
`To determine whether the RTA conjugation affected binding
`affinity, Scatchard analysis was performed for the most effective
`(TA1) and least effective (BD5 and RC6) immunotoxins before and
`after conjugation to RTA (Table II). The relative binding affinities
`for TA1 and TA1-RTA, BD5 and BD5-RTA were identical before
`and after conjugation. However, for RC6 the conjugation process
`decreased the relative binding affinity by about 10-fold. The
`binding sites/cell were within a 2-fold range for both RC6 and
`RC6-RTA.
`To establish that each immunotoxin inhibited protein synthesis at
`a comparable level and that the ricin A chain remained active after
`the conjugation process, immunotoxins were tested in a rabbit
`reticulocyte lysate in vitro translation system (Table III).
`
`Internalization levels of different anti-p185 antibodies and
`immunotoxins are similar
`Ricin-conjugated and unconjugated forms of TA1, BD5 and RC6
`were tested for efficiency of internalization in SKBr3 cells. After 1
`hr of incubation at 37°C with antibodies, followed by protease
`treatment to remove cell surface-bound antibodies, the amount of
`
`Antibodies/immunotoxins
`
`% internalization2
`
`TAI
`TAI-RTA
`BD5
`BD5-RTA
`RC6
`RC6-RTA
`
`13 6 3
`7 6 2
`12 6 4
`11 6 1
`8 6 1
`8 6 1
`
`1The percentage of internalization after one hr incubation with
`antibody/immunotoxin at 37°C was calculated as described in Material
`and Methods.–2Mean 6 SE.
`
`cell-associated radioactivity representing internalized antibodies
`was determined. Results are presented in Table IV. RTA conjuga-
`tion did not affect the amount of internalization observed when
`immunotoxins were compared to unconjugated antibody. All 3
`immunotoxins, each recognizing a different epitope of p185
`protein, produced similar levels of internalization.
`
`Position and orientation of the relative epitope map on the
`extracellular domain of p185
`Previous competitive inhibition studies with radiolabeled antibod-
`ies have defined an epitope map for the extracellular domain in
`which most of the epitopes were arranged in a linear array (Xu et
`al., 1993). The results, however, could not indicate the location of
`epitopes with respect to the peptide sequence of p185. To determine
`the position and orientation of the relative epitope map on the p185
`sequence, deletion mutants of c-erbB-2 were prepared, transfected
`into NIH3T3 cells and evaluated for expression of the different
`p185 epitopes. NIH3T3 cells that expressed full-length p185
`(17313) (McKenzie et al., 1989) and NIH3T3 cells that expressed a
`deletion construct lacking most of the extracellular domain (LTR-1/
`DNerbB-2) (Di Fiore et al., 1987) were obtained for study (Fig. 1).
`In the B2mro construct, most of the intracellular domain was
`deleted. In the Hinc 6–17, Eco 2–20 and Bgl 5–33 constructs,
`portions of the extracellular domain were deleted. The deletions in
`the Hinc 6–17 (codons 242–529) and Eco 2–20 (codons 78–241)
`constructs share a common border between codons 241 and 242.
`The deletion in the Bgl 5–33 (codons 149–370) construct spans the
`common border shared by the Hinc 6–17 and Eco 2–20 constructs.
`NIH3T3 transfectants that contained deletions in the extracellu-
`lar domain were screened for expression of the construct by
`immunohistochemical staining with antibodies that recognized the
`intracellular domain of p185. Only those transfectants that were
`positive for expression were further evaluated in indirect live cell
`radioimmunoassays with antibodies that recognized different epi-
`topes on the extracellular domain of p185. The 17313 cells that
`expressed full-length p185 and the extracellular domain epitopes
`were used as a positive control. As expected, all 10 of the anti-p185
`antibodies bound to 17313 cells (Fig. 3). Deletion of the intracellu-
`lar domain had no effect on the binding of antibodies to the
`extracellular domain. The Hinc 6–17 construct retained binding for
`the 454C11, PB3, NB3 and RC6 antibodies. The complementary
`Eco 2–20 construct retained binding for the RC6, ID5, BD5, 741F8
`and 520C9 antibodies. Although the Bgl 5–33 construct was
`expressed based on binding of antibodies to the intracellular
`domain, it did not retain any of the epitopes recognized by the 10
`antibodies to the extracellular domain. Since the Eco 2–20 deletion
`was N-terminal to the Hinc 6–17 deletion, this oriented the PB3,
`454C11 and NB3 epitopes N-terminal to the ID5, BD5, 741F8 and
`520C9 epitopes.
`When epitopes were retained by a construct, both the conforma-
`tion and sequence information for the epitope were maintained.
`When epitopes were lost it is possible that the conformation was
`destroyed or that the sequence was deleted. In the case of the PB3
`and NB3 epitopes, antibody binding did not require native p185
`conformation since these antibodies bound p185 after electrophore-
`
`IMMUNOGEN 2038, pg. 5
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`530
`
`BOYER ET AL.
`
`immunotoxin did have approximately 10-fold lower relative bind-
`ing affinity compared with the unconjugated RC6 antibody. Whether
`the lower affinity accounted entirely for the decreased cytotoxicity
`of RC6-RTA is not clear, since both unconjugated RC6 and
`RC6-RTA produced similar amounts of internalization.
`Different anti-HER-2 antibodies have been reported by Tecce et
`al. (1993) to have different activities as immunotoxins. Five
`anti-HER-2 antibodies conjugated to a recombinant form of
`Pseudomonas exotoxin exerted different levels of activity (Batra et
`al., 1992). Two of these antibodies, e23 and e21, apparently
`recognized different epitopes, since together, but not individually,
`they inhibited growth of gastric cancer cells in vitro and in vivo
`(Kasprzyck et al., 1992). In another experiment, 2 different
`anti-HER-2 antibodies that recognize different epitopes and did not
`inhibit growth had similar cytotoxic activities when conjugated to
`saporin (Tecce et al., 1993).
`It is possible that functional properties of the different antibod-
`ies, such as inhibiting ligand binding, might affect their cytotoxic
`potency as immunotoxins. Unconjugated anti-HER-2 antibodies
`exert diverse cellular activities depending on their binding site on
`the protein, but independently of inhibiting ligand binding (Lewis-
`Phillips et al., 1998). Both BD5 and ID5 antibodies competitively
`inhibit the binding of the gp30 ligand (Xu et al., 1993). Our data
`demonstrate that ID5-RTA is a more potent immunotoxin than
`BD5-RTA for SKBr3 cells that secrete p30 ligand. This sug