29-40 30/5/07 11:37 Page 29
`
`INTERNATIONAL JOURNAL OF ONCOLOGY 31: 29-40, 2007
`
`29
`
`Characterization of novel murine anti-CD20 monoclonal
`antibodies and their comparison to 2B8 and c2B8 (rituximab)
`
`MICHIO NISHIDA1, SADAKAZU USUDA2, MASATO OKABE2,3, HIROKO MIYAKODA3,
`MIDORI KOMATSU4, HIROSHI HANAOKA5, KEISUKE TESHIGAWARA1 and OTSURA NIWA1
`
`1Late Effects Studies, Radiation Biology Center, Kyoto University, Yoshida-Konoe-cho, Sakyo-ku, Kyoto 606-8501;
`2BioMedics Japan, Inc, 1-1-10 Koraku, Bunkyo, Tokyo 112-0004; 3Bacteriology Division, Faculty of Medicine,
`Tottori University, 86 Nishi-machi, Yonago 683-8503; 4Digestive Surgery, Osaka City University Graduate
`School of Medicine, 1-4-3 Asahi-cho, Abeno-ku, Osaka 545-8586; 5Bioimaging Information Analysis,
`Gunma University, Graduate School of Medicine, 3-39-22 Showa-cho, Maebashi 371-8511, Japan
`
`Received March 2, 2007; Accepted April 27, 2007
`
`Abstract. Rituximab is the first anti-cancer antibody approved
`by the FDA for the treatment of B-cell non-Hodgkin lymphoma
`(B-NHL), alone or in combination with chemotherapeutic
`drugs. Further, rituximab is now being examined in a variety
`of CD20+ neoplastic diseases as well as B-cell-induced
`autoimmune diseases. The clinical response to rituximab is
`significant, resulting not only in tumor regression but also
`prolongation of survival. However, a subset of patients does
`not initially respond to rituximab or develops resistance to its
`further treatment. Therefore, alternative therapies for these
`patients are strongly desired. Rituximab activity has been
`thought to be by antibody-dependent cellular cytotoxicity,
`complement-dependent cytotoxicity and apoptosis, and
`studies in model systems established the role of rituximab in
`cell signaling-induced perturbation of anti-apoptotic survival
`pathways, suggesting that the patients unresponsive to
`rituximab may be overcome with other CD20 antibodies
`with different activities. This study investigated eight novel
`murine antibodies directed against CD20 for their physical
`and biological properties in comparison with 2B8 and c2B8
`(rituximab). These antibodies were derived by various antigenic
`and immunization procedures and selected for CD20 activity.
`Analysis of these antibodies revealed that they all bound to
`various B-cell lines and CD20-transfected CHO cells. Six of
`the eight antibodies shared similar variable-region amino acid
`sequences that were also shared by 2B8 while two monoclonal
`antibodies did not. Of them, 1K1791 has a distinct heavy chain
`and both 1K1791 and 1K1782 have distinct light chains. Not
`all of the antibodies inhibited cell growth and only two anti-
`
`_________________________________________
`
`Correspondence to: Michio Nishida, Late Effects Studies,
`Radiation Biology Center, Kyoto University, Yoshida-Konoe-cho,
`Sakyo-ku, Kyoto 606-8501, Japan
`E-mail: minishida@aol.com
`
`Key words: anti-CD20 monoclonal antibody, apoptosis, caspase,
`epitope specificity
`
`bodies reacted with fixed GST-CD20 recombinant fusion
`protein. Noteworthy, 1K1791 was found to inhibit cell pro-
`liferation and also induced caspase-independent apoptosis in
`the absence of cross-linker. These findings identified new
`antibodies with properties and epitope specificities different
`from 2B8. The potential clinical application of such antibodies
`in the treatment of B-NHL and rituximab-resistant B-NHL is
`discussed.
`
`Introduction
`
`Since the introduction of the first monoclonal antibody (mAb),
`rituximab, for the treatment of B-cell non-Hodgkin lymphoma
`(B-NHL), several monoclonal antibodies have been approved
`for many neoplastic and non-neoplastic diseases (1,2). How-
`ever, while such antibodies show therapeutic effectiveness,
`their mechanism of action in vivo remains elusive. Further, it
`is not known what are the best regimens and schedules under
`which such antibodies are best suited for a particular patient.
`In addition, treatment with such antibodies results only in a
`subset of patients responding and some patients who initially
`respond also develop resistance to further treatment. Therefore,
`there is a need to unravel the in vivo mechanisms of action as
`well as develop new classes of antibodies with different
`activities and clinical responses.
`The CD20 molecule is a member of the MS4A family of
`proteins. It spans the plasma membrane four times and both
`C- and N-termini are contained within the cell (3-6). The
`precise function of CD20 is unknown and it has been reported
`to act as an ion channel expressed in the plasma membrane of
`normal and malignant B cells. Liang et al have reported that
`CD20 may operate as a calcium storage channel facilitating
`entry of intracellular calcium following BCR-induced emptying
`of intracellular calcium (4). The importance of CD20 as a
`target for mAb immunotherapy is irrefutable, and anti-CD20
`monoclonal antibodies appear to be ideal for B-cell diseases.
`It is highly expressed on the plasma membrane of almost all
`plasma B cells but not on hematological stem cells; it is not
`shed from the surface after antibody binding and is not shed
`into the circulation (5).
`
`Ex. 2034-0001
`
`

`
`29-40 30/5/07 11:37 Page 30
`
`30
`
`NISHIDA et al: NOVEL MURINE ANTI-CD20 MONOCLONAL ANTIBODIES
`
`It has been reported that rituximab can mediate antibody-
`dependent cellular cytotoxicity (ADCC), complement-
`dependent cytotoxicity (CDC) and apoptosis both in vitro
`and in vivo, and the target epitope is critical in determining
`which of these various mechanisms predominates (7,8). In
`the case of anti-CD20, evidence suggests that Fc/Fc-receptor
`(FcR) interactions are critical, as determined in both animal
`models and humans. However, determining whether such
`interactions are required for classical ADCC-mediated by NK
`or myeloid cells or whether they provide cross-linking which
`promotes apoptosis has been difficult to resolve (9,10). The
`role of complement in the depletion of malignant cells is
`less convincing, and a number of anti-tumor mAbs appear to
`operate in the absence of lytic complement (11-13). A strong
`correlation between the level of CD20 expression and thera-
`peutic outcome for rituximab has been reported, although
`there is also contrary evidence where no correlation was
`found (14,15). Chan et al have reported a CD20 mAb which
`is potent in CDC and less effective in apoptosis, whereas a
`different antibody was ineffective in CDC with potent induction
`of apoptosis (16). Both were equally effective in ADCC.
`Rituximab is currently being used in the management of NHL
`patients as a single agent or in combination with CHOP. It also
`shows clinical response in patients with chronic lymphocytic
`leukemia (17). Rituximab is also used in autoimmune disease
`such as rheumatoid arthritis and has recently been approved for
`treatment of patients with moderate to advanced rheumatoid
`disease (18).
`Following binding of anti-CD20 mAb to cells recruitment
`of effector cells for ADCC (antibody-dependent cellular
`cytotoxicity) and CDC (complement-dependent cytotoxicity)
`occurs. In addition, when CD20 is engaged by mAb, it can
`trigger transmembrane signaling directly and inhibit cell growth
`and trigger cell death in certain tumors. Further, it has also
`been shown to sensitize tumor cells to both chemotherapy
`and immunotherapy (6). Different anti-CD20 mAbs have been
`shown to have different properties and epitope specificities,
`and mediate differential effects on CDC and cell death. All of
`the monoclonal antibodies described to date recognize the
`extracellular loop and partially or completely cross-block
`each other's binding (19-21).
`In the current study, we have developed a number of murine
`anti-CD20 mAbs and compared their activities to the published
`antibodies 1F5, 2B8 and c2B8. We have identified new mAbs
`with different activities and binding specificities.
`
`Materials and methods
`
`mice (purchased from CLEA Japan, Tokyo, Japan) were
`immunized with various combinations of Raji cells, CCRF-SB
`cells, recombinant fusion protein of CD20 extracellular domain
`with glutathione S-transferase (GST) and CD20 DNA trans-
`fected CHO cells (23).
`The murine anti-CD20 mAb 2B8 and the chimeric anti-
`CD20 mAb c2B8 (rituximab) were obtained from Zenyaku
`Kogyo (Tokyo, Japan). The murine anti-CD20 antibody 1F5,
`normal murine IgG1, IgG2a and IgG2b were purchased from
`Dako Japan (Kyoto, Japan), and murine anti-human CD3
`mAb was from BD Biosciences (San Jose, CA).
`
`Flow cytometry. Various mAbs and control immunoglobulin
`(Ig) isotypes were used to determine their binding properties
`on the cell surface of CD20+ cell lines or CD20- cells. The cells
`(2x106 cells) were suspended in RPMI-1640 culture medium
`(Sigma Chemical, St Louis, MO) containing 10% fetal bovine
`serum (FBS) (ICN Biochemicals, Costa Mesa, CA) and were
`centrifuged at 1300 x rpm for 3 min. The supernatants were
`discarded and the pellets resuspended in 550 μl phosphate
`buffered saline (PBS) containing 4% FBS. Aliquots of 50 μl
`PBS containing 4% FBS each were placed in Eppendorf tubes
`and 5 μl (10 μg/ml) of each antibody or control IgG were
`added to the tube and incubated for 30 min at 4˚C. The tubes
`were centrifuged and the pellets washed twice and then
`resuspended into 50 μl PBS containing 4% FBS and 5 μl
`FITC-conjugated F(ab')2 goat anti-mouse Ig (Dako, Japan).
`The tubes were left for 20 min at 4˚C, washed twice and the
`pellets were resuspended in 12x75-mm tubes (BD Falcon,
`BD Biosciences). The flow data were then analyzed by FACS
`Calibur and by the Cell Quest software (BD Biosciences).
`
`Inhibition of cell growth. Raji cells (5x104 cells/ml) in RPMI-
`1640 serum-free medium were assessed for cell growth
`inhibition in the presence of various antibodies or control
`IgGs. The cells (5x103 cells/well) were added into 96-well
`plates and cultured at 37˚C in a 5% CO2 incubator for 24 h.
`Thereafter different concentrations of antibody (0.01-1.0 μg/ml)
`were added and incubated for 24-72 h. At the indicated times of
`incubation, 10 μl/well of luminescent reagent (Cell Count Kit-
`8, Dojundo Laboratories, Kumamoto, Japan) was added into
`each well and incubated for an additional 4 h. The plates were
`then read in a microplate reader (Hitachi High-Technologies,
`Tokyo, Japan) and the absorbance at 492 nm was recorded.
`The percent of cell growth in the presence of each antibody
`(0.01-1.0 μg/ml) compared to control IgGs and % inhibition
`were determined.
`
`Cells and antibodies. The CD20+ cell lines, Raji (Burkitt's
`lymphoma), CCRF-SB (acute lymphoblastic leukemia) and the
`CD20- cell line Jurkat (T cell leukemia), were obtained from
`the Riken Bioresource Center, Japanese Collection of Research
`Bioresources (JCRB; Tsukuba, Japan) and the American Type
`Culture Collection (ATCC; Manassas, VA). CHO DG44 was
`obtained from Invitrogen Japan (Tokyo, Japan) (22). The full
`CD20 DNA-transfected CHO cell line (CD20+ CHO) was
`collaboratively developed by Tottori University (Yonago,
`Japan) and BioMedics Japan, Inc. (Tokyo, Japan).
`The murine 1K mAbs (Table I) specifically binding to
`human CD20 were provided by BioMedics Japan. BALB/c
`
`Affinity measurement by Scatchard analysis. Radiolabeling
`of antibodies by direct radio-iodination was performed after
`the chloramine T method (24). 2 μl of Na 125I (740 kBq) was
`added to 100-μl aliquots of antibody (0.1 mg/ml) in 0.3 M
`phosphate buffer (pH 7.4) and chloramine T (1 μg in 3 μl),
`freshly prepared in the same buffer, was added thereafter.
`After incubation of the mixture for 5 min at room temperature,
`125I labeled antibody was purified by the Bio-Spin 6 column
`(Wako Pure Chemical, Osaka, Japan).
`Binding affinity was determined as follows. Buffer solution
`(100 μl) was added to 1x106 Raji cells and then mixed together.
`Each 100 μl of 125I labeled antibody solution (containing
`
`Ex. 2034-0002
`
`

`
`29-40 30/5/07 11:37 Page 31
`
`INTERNATIONAL JOURNAL OF ONCOLOGY 31: 29-40, 2007
`
`31
`
`Table I. Properties of monoclonal antibodies used in this study.
`–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
`mAba
`Isotype
`Immunization
`Source
`Kd
`Bmax
`–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
`IgG2b, κ
`1K0924
`2.23 nM
`0.7x106
`CCRF-SB and GST-CD20
`BioMedics Japan
`IgG1, κ
`1K1228
`1.26 nM
`1.6x106
`CD20+CHO and GST-CD20
`BioMedics Japan
`IgG1, κ
`1K1402
`1.25 nM
`1.7x106
`CCRF-SB and CD20+ CHO
`BioMedics Japan
`IgG1, κ
`1K1422
`2.07 nM
`1.3x106
`CCRF-SB and CD20+ CHO
`BioMedics Japan
`IgG2a, κ
`1.70 nM
`0.7x106
`CD20+ CHO and Raji
`BioMedics Japan
`1K1712
`IgG2b, κ
`1K1736
`1.24 nM
`1.7x106
`CD20+ CHO and Raji
`BioMedics Japan
`IgG1, κ
`1K1782
`NT
`NT
`CD20+ CHO and Raji
`BioMedics Japan
`IgG1, κ
`1K1791
`3.61 nM
`1.5x106
`CD20+ CHO and Raji
`BioMedics Japan
`IgG2a, κ
`1.41 nM
`0.8x106
`NA
`ATCC
`1F5
`IgG1, κ
`2B8
`NT
`NT
`CCRF-SB
`Zenyaku Kogyo
`IgG1, κ
`c2B8 (rituximab)
`1.26 nM
`1.3x106
`Chimerized mAb of 2B8
`Zenyaku Kogyo
`–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
`NT, not tested; NA, information not available. a1K series of monoclonal antibodies were newly developed by BioMedics Japan. 2B8 and
`c2B8 that were developed by IDEC Pharmaceuticals (San Diego, CA) were obtained through Zenyaku Kogyo.
`–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
`
`0.01 μg antibody) and different quantities of non-labeled
`antibody, 0.01, 0.03, 0.1, 0.3, 1.0, 3 or 10 μg, were added to
`100 μl Raji cell suspension and then reacted for 1 h at room
`temperature. The supernatant from the cell suspension was
`removed after centrifugation at 1300 rpm for 3 min, and the
`radioactivity of the cell fraction was measured. The radio-
`activity of 125I-labeled antibody bound to cells was counted
`and compared to the initial radioactivity. The binding constants,
`either Kd (dissociation constant) or Ka (association constant),
`were determined by Scatchard analysis by plotting specific
`bound/free ratio (Y axis) vs. bound antibody concentration
`(X axis). Simultaneously, Bmax (receptor density) was
`determined by fitting to the total binding at the saturated
`concentration of non-labeled antibody.
`
`each well on 96-well PLL plates and immobilized. Blocking
`treatment was performed at 37˚C for 1 h. Rituximab solution
`(40 and 80 ng/ml) and 2.0, 0.4 and 0.08 μg/ml of diluted test
`antibodies as well as 2B8 were prepared. Each combination
`of diluted rituximab and test antibody was added with CD20+
`CHO cells on a cell ELISA plate. The first set of reactions
`used 60 μl/well of antibody at 37˚C for 4 h and the second
`set of reactions used diluted solution (x2000) of HRP-labeled
`anti-human IgG (Jackson Lab) at 37˚C for 1 h, respectively.
`The color reaction was developed for 20 min at 25˚C under
`dark and static conditions. H2O2 + OPD (100 μl/well) was
`used for the luminescent liquid and 50 μl/well of 4N H2SO4
`for the stop solution, and then the absorbance at 492 nm was
`measured.
`
`ELISA with GST-CD20. Glutathione S-transferase (GST)
`fused with the extracellular domain of CD20 was obtained
`from the Biochemistry Division, Faculty of Medicine, Tottori
`University (Yonago, Japan). This fusion protein was solubilized
`using 6 M sodium hydroxide solution and immobilized on a
`96-well PLL plate and washed with a solution containing
`150 mM NaCl, 0.5% Tween-20 and 0.1% NaN3. Blocking
`treatment was applied at 37˚C for 4 h using the dilution buffer
`containing 0.2% gelatin, 0.5% bovine serum albumin (BSA)
`and 0.01% thimerosal. Antibody sample [100 μl (1000, 316,
`100, 32, 10 or 3 ng/ml)] was added to the wells and left to
`react at 37˚C for 4 h (1st reaction). 2B8 was used as a positive
`control and solution without antibody was used as a negative
`control. The plate was washed and then 100 μl solution (diluted
`to 1:4000) of horseradish peroxidase (HRP)-labeled rabbit
`anti-mouse IgG (Jackson Lab, ME) was added to each well,
`and incubated at 37˚C for 1 h (2nd reaction). The plate was
`washed again and 100 μl luminescent liquid was added to
`each well and the reaction stopped after 30 min by adding 50 μl
`of 4N H2SO4. The plates were read for the absorbance at
`492 nm.
`
`Competitive binding assay using the CD20+ CHO ELISA. Of
`CD20+ CHO cells, 100 μl (5x105 cells/ml) were added to
`
`Phylogenetic tree analysis of anti-CD20 mAb variable-region
`amino acid sequences. The relative differences of amino acid
`sequences of variable regions between a number of anti-CD20
`mAbs and 2B8 were mapped and displayed by applying a
`neighboring-joining method on phylogenetic trees (25). Heavy
`and light chain variable-region sequences of anti-CD20 mAbs
`were obtained from several information sources. Those of
`2B8, 2H7 and 1F5 were obtained from GenBank (National
`Center for Biotechnology Information, USA), 9C10, 12E11
`and 1K mAbs were derived from BioMedics Japan. These
`data were analyzed by a computer at the National Institute of
`Genetics (Mishima, Japan).
`
`Cell death and apoptosis. Apoptosis and necrosis induced
`by the various mAbs were measured by flow cytometry with
`Annexin V/FITC, PI (propidium iodine) staining (Annexin V/
`FITC apoptosis detection kit, BD Biosciences). 2B8 was used
`as positive control and murine anti-CD3 mAb (BioLegend,
`San Diego, CA) as the negative control (note: the amount of
`1K1782 was not enough to complete the study).
`
`Analysis of caspase activation. For titrating rituximab, 100 μl
`of RPMI-1640 medium containing 10% heat-inactivated FBS
`and penicillin/streptomycin (Invitrogen, Carlsbad, CA) were
`
`Ex. 2034-0003
`
`

`
`29-40 30/5/07 11:37 Page 32
`
`32
`
`NISHIDA et al: NOVEL MURINE ANTI-CD20 MONOCLONAL ANTIBODIES
`
`Figure 1. Phylogenetic tree analysis of amino acid sequence of anti-CD20 antibody variable regions. The relative differences of amino acid sequences of
`variable regions between a number of anti-CD20 mAbs and 2B8 were phylogenetically mapped employing a neighboring-joining method. Both the heavy and
`light chain sequences of most antibodies were very close to each other. Only the monoclonal antibody 1K1791 had a distinct heavy chain variable region and
`both 1K1791 and 1K1782 had light chains different from most other anti-CD20 mAbs.
`
`added to each well of a 96-well TC plate (E&K Scientific
`Products, Santa Clara, CA). Raji cells [100 μl (2x105 cells/ml)]
`were added to the wells of the first row and then 100 μl
`serially diluted down each column. Rows were divided into
`those with and those without antibody. For the antibody
`containing wells, medium containing 10 μg/ml of both
`rituximab and goat anti-human IgG were added and the plate
`was incubated in a humidified atmosphere of 5% CO2 at
`37˚C for 24 h. Of either caspase-3/7 substrate (Promega, WI)
`or caspase-9 substrate (Promega), 100 μl was then added to
`the appropriate rows and the plates incubated on a shaker
`(based on the Promega technical bulletin of Caspase-Glo™
`3/7 Assay System: TB323, Caspase-Glo™ 9 Assay System).
`The luminescence was read after 1 and 2 h by a Wallac Victor2
`plate reader (Perkin-Elmer, Waltham, MA). Duplicate control
`wells containing 20 μM zVAD-fmk as well as either caspase-
`3/7 or caspase-9 substrate and 5000 or 1250 cells, respectively,
`were included on the plate.
`For 1K mAbs testing, Raji cells (5x103 cells/well) in 85 μl
`of medium were placed in each well of a 96-well TC flat
`bottom plate. Half the rows contained 10 μM zVAD-fmk and
`
`half medium alone. Of test antibody, 5 μg/ml was added to
`each well followed by 5 μg/ml of goat anti-murine IgG.
`Medium was added to all wells to bring the final volume up
`to 100 μl. Appropriate control wells with or without test
`antibody or using human IgG or murine IgG were included.
`Ionomycin (Calbiochem™ EMD Biosciences, San Diego,
`CA) in a concentration of 1 μM was used as a positive control.
`Cells were incubated in a humidified atmosphere of 5% CO2
`at 37˚C for 12 h and an equal volume (100 μl) of caspase-3/7
`substrate was added. The plate was agitated at room temp-
`erature for 50 min and the luminescence read by a Wallac
`Victor2 plate reader.
`
`Results
`
`Characterization of novel murine anti-CD20 mAbs. Different
`protocols were used to generate murine mAbs directed against
`human CD20. Cell lines expressing CD20 or CD20-transfected
`CHO cells were used for priming and boosting. In addition,
`recombinant GST-CD20 fusion protein was used for boosting.
`Several mAbs were selected for further analysis and were
`
`Ex. 2034-0004
`
`

`
`29-40 30/5/07 11:37 Page 33
`
`INTERNATIONAL JOURNAL OF ONCOLOGY 31: 29-40, 2007
`
`33
`
`FL1
`
`Table II. Mean fluorescence intensity of antibody binding to various cells.a
`–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
`Raji
`CCRF-SB
`CD20+ CHO
`CD20- cells
`––––––––––––––––––
`–––––––––––––––––––
`–––––––––––––––––––
`––––––––––––––––
`MFI
`S.D.
`MFI
`S.D.
`MFI
`S.D.
`MFI
`S.D.
`(average)
`(average)
`(average)
`(average)
`–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
`Rituximab
`7.9x102
`1.3x102
`1.0x103
`4.4x102
`3.8x103
`8.8x102
`1.7
`0.2
`1K mAbs
`6.7x102
`2.1x102
`7.8x102
`4.3x102
`3.1x103
`5.8x102
`1.6
`0.2
`–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
`aThe mean fluorescence intensity (MFI) of antibody binding to different CD20+ CHO cells was determined through integration of the
`fluorescence histogram of each antibody as in Fig. 2. The FITC intensities detected by FL1 were averaged and compared among those of
`different CD20+ CHO cells. Untransfected CHO DG44 or Jurkat cells were tested as negative controls. The results of the 1K mAbs were
`pooled to show a broad range of antibody response to cells rather than response by individual antibodies, such as rituximab.
`–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
`
`given the nomenclature 1K followed by a number, and were
`then compared to murine mAbs 1F5 and 2B8, as well as the
`chimeric anti-human mAb c2B8 (rituximab). Eight mAbs in
`the 1K series were selected for examination in this study. The
`isotypes of the mAbs consisted of IgG1, IgG2a, and IgG2b
`all with a κ light chain (Table I). Affinity measurements were
`assessed by the use of 125I-radiolabeled antibody binding to
`CD20+ Raji cells and its displacement by different concen-
`trations of non-radiolabeled antibodies, as described in
`methods. The dissociation constant (Kd) was determined for
`each antibody and the values are listed in Table I. There were
`no significant differences in the Kd values among the various
`1K mAbs as well as among the other mAbs 1F5, 2B8 and
`c2B8. Receptor density analysis was also examined to
`determine the Bmax values. As shown in Table I, the Bmax values
`of all the antibodies tested ranged from 0.7-1.7x10-6. These
`findings demonstrate that, following the various protocols
`used for immunization, several novel mAbs (1K series)
`directed against human CD20 were generated with different
`isotypes. These antibodies exhibited similar Kd and Bmax
`values compared to both murine 1F5 and 2B8 as well as the
`chimeric anti-human CD20, c2B8 (rituximab), approved by
`FDA. The findings with the 1K mAbs were further explored
`for other properties that might distinguish them from the known
`mAbs tested.
`
`Analysis of amino acid sequence of both light and heavy
`chains of the anti-CD20 variable regions of the different 1K
`mAbs and the published mAbs: homologies for all except for
`1K1782 and 1K1791. The variable regions of both heavy and
`light chains that are involved in antigen binding determine
`both specificity and affinity. Thus, we established the amino
`acid sequences of the anti-CD20 variable regions of the various
`1K mAbs and determined homology as well as non-homology
`among these and published anti-CD20 mAbs including 2B8,
`1F5 and 2H7. The data were topologically mapped by tree
`analysis as shown in Fig. 1. In this figure, it is noted that all 1K
`mAbs and the published anti-CD20 mAbs have almost identical
`amino acid sequences of the heavy chain-V region except
`1K1791 (Fig. 1, top). Analysis of the light chain sequences
`revealed that all 1K mAbs share similar sequences of the κ
`light chain-V region except 1K1782 and 1K1791. The
`differences of the V regions of both 1K1782 and 1K1791
`
`suggest that these two antibodies may exhibit different
`properties as compared with the remaining 1K mAbs or the
`published anti-CD20 mAbs. Likewise, the homology observed
`between the six mAbs suggests that these antibodies may
`exert similar functions. However, the epitope specificities of
`these various 1K mAbs could not be assigned based on the
`homology and differences in the amino acid sequence of the
`variable region or the heavy and light chains.
`
`Binding of the various 1K mAbs to CD20-expressing cell
`lines as expressed by flow cytometry. The cell surface binding
`properties of the various mAbs to CD20-expressing cells
`were determined by flow cytometry as described in Materials
`and methods. Raji, CCRF-SB, and CD20+ CHO cells were
`used as CD20+ cells and CHO DG44 (CD20-) or Jurkat cells
`were used as a negative control. Control normal murine Ig
`was used with the corresponding isotype to the test mAb. The
`cells were treated with an excess amount of antibody to bind,
`washed and treated with FITC-conjugated F(ab')2 goat anti-
`mouse Ig and processed as described in Materials and methods
`for flow analysis. The data are represented in histograms
`(Fig. 2). For comparison, the 2B8 mAb staining is depicted in
`purple color, the isotype controls are depicted by green dotted
`lines. Clearly, there was no staining of the CHO DG44 and
`Jurkat cells demonstrating the specificity of the 1K mAbs.
`The histogram analyses show that all 1K mAbs stained Raji
`cells with similar intensity. With CCRF-SB, 1K mAbs stained
`the cells with higher intensity than with Raji and there were
`some differences such as that 1K1791 and 1K1782 showed
`less staining compared to 2B8 and other 1K mAbs. Analysis
`of the CD20-transfected CHO cells showed that all 1K mAbs
`stained the cells with significantly higher intensity as compared
`to Raji and CCRF-SB. The high intensity staining of CD20+
`CHO cells reflects the overexpression of cell surface CD20
`compared to non-transfected Raji and CCRF-SB (Table II).
`In addition to flow cytometry, the binding of the various
`mAbs to the CD20+ CHO cells was also analyzed by ELISA,
`as described in Materials and methods. Different concentrations
`of the antibodies were used, their binding to CD20+ CHO cells
`determined and the data were analyzed (Fig. 3). The binding
`was found to be dependent on the concentration of the antibody
`used. At a low concentration of 3 ng/ml, there was little binding
`of 2B8 and 1K mAbs except for 1K1712 that was relatively
`
`Ex. 2034-0005
`
`

`
`29-40 30/5/07 11:37 Page 34
`
`34
`
`NISHIDA et al: NOVEL MURINE ANTI-CD20 MONOCLONAL ANTIBODIES
`
`Figure 2. Flow cytometry analysis of the various monoclonal antibodies. The FITC fluorescence intensity was detected by FL1 in flow cytometry analysis.
`Rows 1-8 are 1K0924, 1K1228, 1K1402, 1K1422, 1K1712, 1K1736, 1K1782 and 1K1791 and columns 1-3 show the reactivity to Raji, CCRF-SB and CD20+
`CHO respectively. The 4th column shows binding to CHO DG44 host (CD20-) (lines 1-4) and Jurkat (lines 5-8). 2B8 staining is depicted in a solid purple
`color as a positive control and its isotype control, IgG1, κ by green dotted line. 1K mAbs are depicted by a pink solid line and isotype control antibodies by an
`aqua dotted line.
`
`higher than others. By increasing the concentration of the
`antibodies, there was increased binding of up to 100-316 ng/ml
`and a plateau was achieved. The binding by the various 1K
`mAbs showed some differences. For example, 1K1712 was
`the most efficient binder followed by 1K1736 and there was
`equal binding by 2B8, 1K0924, 1K1782 and 1K1402. Both
`1K1791 and 1K1422 were the poorest binders. The binding
`
`was specific as control IgGs did not show any activity. These
`findings suggest that the various mAbs may have different
`activity on CD20+ overexpressing cells.
`
`Inhibition of Raji cell growth by 1K mAbs. The previous
`findings demonstrated the ability of various 1K mAbs to
`bind cell surface expression on various CD20+ cell lines. We
`
`Ex. 2034-0006
`
`

`
`29-40 30/5/07 11:37 Page 35
`
`INTERNATIONAL JOURNAL OF ONCOLOGY 31: 29-40, 2007
`
`35
`
`Figure 3. Analysis of antibody binding to CD20+ CHO by whole cell
`ELISA. Antibody reactivity to CD20+ CHO fixed on a PLL plate was tested
`as described in Materials and methods.
`
`Figure 5. Competitive inhibition of binding of c2B8 on Raji by the various
`1K and 2B8 monoclonal antibodies. Two concentrations of c2B8, 40 and
`80 ng/ml, were used and different concentrations of 1K mAbs and 2B8 were
`used for competitive binding as described in Materials and methods.
`
`Figure 4. Inhibition of Raji cell growth by the various monoclonal antibodies.
`The inhibitory effects of the 1K series of mAbs and 2B8 at two different
`concentrations (0.1 or 1.0 μg/ml) on Raji cells were examined over 24-72 h.
`The percent inhibition of cell growth was determined as cell growth without
`mAbs (100%) minus cell growth rate (%) in the presence of mAb, where
`cell growth was determined by the luminescence.
`
`examined the effect of these antibodies on cell growth of Raji
`cells in culture (Fig. 4). Raji cells were treated with various
`concentrations of 1K mAbs and cultured for up to 72 h.
`The cells were harvested at different times and cell growth
`was determined by the luminescent method as described in
`Materials and methods. When 0.1 μg/ml mAb was used,
`there was significant inhibition of cell growth by 1K1791,
`1K0924, 1K1422 and 2B8 at both 48 and 72 h of culture. The
`remaining antibodies did not exert any significant inhibitory
`effect on the growth of Raji cells. When a higher concentration
`of antibody was used (1.0 μg/ml), there was a more pronounced
`inhibition. The mAbs 1K1791, 1K1422, 2B8 and 1K0924,
`showed a time-dependent inhibition of cell growth, though
`1K0924 showed less inhibition than the other three mAbs.
`Even at a higher concentration of 1.0 μg/ml, 1K1402, and
`1K1782 did not result in any inhibitory activity of Raji cell
`growth. These findings demonstrate that three of the 1K
`mAbs behave like 2B8 in inhibiting Raji cell growth in a
`time- and concentration-dependent manner. Further, the studies
`demonstrate that, although the non-inhibiting 1K mAbs bind
`to CD20 on Raji cells with the same intensity as the three
`inhibitory 1K mAbs, the mere binding is not sufficient to exert
`
`Figure 6. Binding of the various 1K mAbs and 2B8 to recombinant fixed
`GST-CD20. The binding of the various mAbs to GST-CD20 was performed
`as described in Materials and methods.
`
`an inhibitory effect of cell growth. It is possible, however,
`that the non-inhibitory activity on Raji cells may not be
`generalized to other cell lines. The findings demonstrating
`that three mAbs can inhibit cell growth suggest that these
`antibodies signal the cells through their interaction with their
`CD20 receptor and affect DNA replication. The inhibitory
`effect of cell growth shown above by the 1K mAbs did not
`discriminate between the cytostatic effect and cytotoxic effect
`and further analyses are shown below.
`
`Epitope specificity of the 1K mAbs as assessed by competitive
`inhibition of binding of 2B8 to CD20+ CHO cells. 2B8 mAb
`was used to genetically engineer the chimeric c2B8 (rituxi-
`mab) that was approved by the FDA in 1997 for the treatment
`of B-NHL. Thus, we examined the relationship between
`the epitope binding specificity of the various 1K mAbs in
`comparison to that of c2B8. Competitive binding was assessed
`by using immobilized CD20+ CHO cells on 96-well plates
`and the luminescent ELISA method was used as described in
`Materials and methods. Two concentrations of c2B8 were
`used, 40 and 80 ng/ml. The findings are presented in Fig. 5.
`At a lower concentration of 40 ng/ml of c2B8 mAb, all of the
`1K mAbs and 2B8 mAb significantly competitively inhibited
`the binding of c2B8 on CD20+ CHO cells. The inhibition
`was dependent on the concentration of the antibody used; a
`
`Ex. 2034-0007
`
`

`
`29-40 30/5/07 11:37 Page 36
`
`36
`
`NISHIDA et al: NOVEL MURINE ANTI-CD20 MONOCLONAL ANTIBODIES
`
`Figure 7. Analysis of cell death mediated by various mAbs on Raji cells.
`Raji cells were treated with the mAb for 24 h (A) and 48 h (B). 2B8 was
`used as a positive control and murine anti-CD3 mAb as the negative control.
`
`concentration of 2 μg/ml was more inhibitory than lower
`concentrations. All of the 1K mAbs showed similar competitive
`inhibition at all 3 antibody concentrations used. Interestingly,
`2B8 competed less strongly than the 1K mAbs for c2B8
`binding. Similar findings to those observed with a c2B8
`concentration of 40 ng/ml were also found when a higher
`concentration of 80 ng/ml c2B8 was used.
`These findings suggest that the a