Vol. I. l.J/9-/J25.
`Nm·t'mher 1995
`
`Clinical Cancer Research t319
`
`Monocyte-mediated Lysis of Acute Myeloid Leukemia Cells
`
`
`in the Presence of the Bispecific Antibody 251 x 22
`(Anti-CD33 x Anti-CD64)1
`
`Jian Chen, Jie-Hua Zhou, and Edward D. Ball2
`Division of Hematology/Bone Marrow Transplantation. University of
`Pittsburgh Medical Center and Pittsburgh Cancer Institute. Pitt>burgh.
`Pennsylvania 15213
`
`humanized anti-C033 and anti-Fc-yRI could have clinical
`application in the treatment of myeloid leukemia, especially
`in the management of minimal residual disease.
`
`ABSTRACT
`lmmunotherapy using bispecific antibodies (BsAb) to
`direct immune effector cells toward target tumor cells has
`been shown to be effective in a number of studies. Several
`immune trigger molecules have been characterized. Among
`them, Fc-yRI appears to play an important role in antibody·
`dependent cellular cytotoxicity. It is expressed mainly on
`monocytes, macrophages, and neutrophils under certain
`clinical situations. The expression of Fc")'RI can be regulated
`by a variety of cytokines, primarily by IFN·'Y· Recent studies
`have shown that granulocyte·colony·stimulating factor
`(G-CSF) and granulocyte-macrophage-colony stimulating
`factor (GM-CSF) can increase the number of the Fc")'RI·
`posit.ive monocytes, increase the expression of Fc")'RI on
`circulating neutrophils after in vivo infusion, and greatly
`enhance the cytotoxic activity of circulating neutrophils.
`CD33 is a glycoprotein expressed on the ceJI surface of
`mature monocytes, myeloid progenitor cells. and myeloid
`leukemic blasts, but not on the earliest hematopoiet.ic pro­
`genitor cells and other normal tissues. Herein, we report the
`construction of a BsAb, 251 x 22, by conjugating an anti·
`CD33 mAb (mAb 251) to an anti-Fc")'RI mAb (mAb 22). The
`BsAb 25 I x 22 is capable of enhancing the cytotoxicity of
`several leukemia cell lines by cytokine-activated monocytes.
`Our data also show that G-CSF- and GM-CSF-stimulated
`monocytes can mediate cytotoxicity of target leukemia cells
`comparable to that of IFN-")'-stimulated monocytes. The
`expression of Fc-yRI on monocytes after 24-h in vitro incu­
`bation with G-CSF and GM-CSF was increased, although
`not significantly. Prolonged incubation of monocytes with
`G-CSF for 48 h significantly increased the Fc-yRI expres­
`sion. Because humanized anti-C033 and anti-Fc-yRI mAb
`are available, and because GM-CSF and G-CSF have been
`used widely for patients after chemotherapy to stimulate the
`recovery of myeloid hematopoiesis, additional clinical devel­
`opment of this project is feasible. A BsAb comprised of
`
`INTRODUCTION
`Specific tumor cell lysis can be achieved by applying
`BsAbs·' to direct immune effector
`cells toward tumor cells.
`This
`novel immunotherapeutic approach has been the focus of nu­
`merous i11 1·irro and i11 1•fro studies ( 1-4). Several immune­
`including FcyRI. FqRll. FqRlll, and
`
`triggering molecules,
`
`
`T -cell receptor-associated molecules. have been shown to play
`
`important roles in various immune responses. FcyRI (CD64) is
`
`
`expressed mainly on 1he surface of monocytes, macrophages.
`(5-7). CD64
`
`and neutrophils under certain clinical conditions
`
`
`expression can be modulated by various cytokines, primarily
`shown that;,, 1-fro admin­
`
`IFN--y (8-10). Previous studies have
`
`istration of G-CSF increases the FcyRI expression of neutro­
`
`
`
`phils (8), whereas administration of GM-CSF in 1-fro can in­
`crease the number of circulating FcyRl-posit ive monocytes
`(10).
`
`GM-CSF has been shown to stimulate neutrophil phago­
`cytosis (6. 8). and to enhance ADCC by monocytes and gran­
`
`ulocytes (5-7). To use monocytes as immune effector cells. the
`
`BsAb should be capable of binding to FcyRI with high affinily
`and to a specific antigen on a tumor cell surface. Such a BsAb
`
`
`could direct monocytes toward tumor cells and elicit the lysis of
`
`
`
`target cells. CD33 is a glycoprotein with res1ric1ed expression on
`
`
`myeloid progenitor cells and leukemic blasts, but not on other
`
`
`normal tissue and early hematopoietic progenitor cells ( 11-13).
`
`
`Clinical trials targeting mAb to CD33 on leukemia cells have
`
`toxicity to normal body tissues (14).
`shown no significant
`previously the synthesis of a BsAb
`We have reported
`
`x anti-FcyRlll). This BsAb was capa­
`251 X 3G8 (anti-CD33
`of nalural killer
`cells against
`
`ble of augmenting the cytotoxicity
`CD33·positive AMI cell lines and fresh cells from patients
`(15).
`In this study. we describe a new BsAb 251 x 22 (anti-CD33
`x
`the cytotoxicity of
`
`anti-Fc-y RI) that can significantly increase
`
`
`
`CD33-positive leukemia cells by cytokine (IFN·'f. G-CSF. and
`
`GM-CSF)-activated monocytes.
`
`MATERIALS AND METHODS
`by E. D. B.
`
`mAbs. mAb 251 (anti·CD33) was developed
`
`
`( 16). Purified 251 and the anti-Fey RI mAb (mAb 22) were
`
`Received 1/11/95: revised 6/13/95: accepted 7/5/95.
`1 This work was supported in part by Grant CA31888 from the National
`�anccr Institute and by the Blood Science Foundation (Pittsburgh. PA).
`·To whom requests for reprint> should be addressed, at the Division of
`Hematology/Bone Marrow Transplantation. University of Pittsburgh
`Medical Center. 200 Lothrop Street. Pittsburgh. PA 15213. Phone: (412)
`648·6413: Fax: (412) 648-6393.
`
`'The abbreviations used arc: BsAb. bispccific antibody: GM-CSF.
`granulocyte-macrophagc colony-stimulating factor: SATA. N-succin·
`imidyl-S-acctylthioacctatc; Sulfo-SMCC. sulf�uccinimidyl 4 (N·malc·
`i"?idom�thyl) cyclohexanc·L·carboxylate: G-CSF, granulocyte colony­
`Sllmulatmg
`factor: AML. acute myeloid
`leukemia: MFI. mean
`nuorcsccncc intensity: ADCC. antibody-dependent cellular cytotoxicity.
`
`Downloaded from
`
`clincancerres.aacrjournals.org
`
`on August 31, 2015. © 1995 American Association for Cancer
`Research.
`
`1 of 8
`
`BI Exhibit 1014
`
`

`

`1320 AML Cell Lysis in the Presence of BsAb 251 X 22
`
`kindly provided by Medarex, Inc. (Annandale, NJ); both are
`murine IgGl mAbs. SCCL-1, an IgG2a murine mAb, directed
`against the transferrin receptor, was used in this study as a
`positive control for ADCC via an unconjugated mAb (17).
`F(ab'h fragment of mAb 25 l was prepared by pepsin digestion
`and purified using a protein A column (Pharmacia Biotech, Inc.,
`Piscataway, NJ).
`Production of BsAb 251 x 22. Conjugation chemicals
`SATA and Sulfo-SMCC were obtained from Pierce Chemical
`Co. (Rockford, IL). Hydroxylamine was purchased from Sigma
`Chemical Co. (St. Louis, MO). SATA (50 mM) was freshly
`prepared in 100% dimethylformamide and Sulfo-SMCC (50
`mM), in PBS (pH 7.4). SATA was mixed with mAb 251 in a
`final molar ratio 10:1. After 1-h reaction at room temperature,
`nonreacted SATA was removed by size-exclusion centrifuga­
`tion through a Centricon 30 apparatus (Amicon, Inc., Beverly,
`MA). The free sulfhydryl group was generated by deacetylation
`with hydroxylamine at room temperature for 2 h. Excess hy­
`droxylamine was removed by a Centricon 30 apparatus. At the
`same time, the Sulfo-SMCC was reacted with mAb 22 in a final
`molar ratio of 20: I at room temperature for 2 h. The maleimide­
`activated mAb22 was separated from unreacted Sulfo-SMCC by
`centrifugation through a Centricon 30 apparatus. The final con­
`jugation was carried out by mixing equal molar amounts of both
`mAbs at room temperature overnight. The concentration of
`BsAbs was determined using a Bio-Rad DC Protein Assay
`(Bio-Rad Laboratories, Richmond, CA). The purity of the
`BsAbs was verified by SOS-PAGE.
`CeU Lines. HL-60 and KG-la cell lines were obtained
`from American Type Culture Collection (Rockville, MD). The
`NB4 cell line was kindly provided by Dr. M. Lanoue (Paris,
`France). Cell lines were maintained in RPMI 1640 medium
`containing 10% FCS, L-glutamine {2 mM), penicillin (100 units/
`ml), and streptomycin (50 µ.g/ml) (GIBCO-BRL, Grand Island,
`NY) at 37°C in a humidified atmosphere with 5% C02• All cell
`contamination
`lines were checked periodically for Mycoplasma
`using the ONA hybridization method (Gen-Probe, San Diego,
`CA). HL-60 and NB4 cells are strongly positive for CD33,
`whereas KG-la cells are negative for CD33.
`of Effector Cells. Monocytes were freshly
`Preparation
`separated from the peripheral blood of normal donors after they
`gave informed consent. After Ficoll-Hypaque {Pharmacia, Pis­
`cataway, NJ) gradient centrifugation, the mononuclear cells
`were collected and washed twice. The cells were incubated with
`mAb OKT3 (American Type Culture Collection) and mAb 6G7
`(anti-CD45RA)4 for I h at 4°C and washed twice. Dynabeads
`M-450 coated with sheep antimouse IgG (DYNAL, Great Neck,
`NY) were added at a final ratio of 20:1 and incubated at 4°C for
`1 h. The lymphocytes were depleted using a handheld magnet.
`The procedure was repeated twice. The purity of monocytes was
`analyzed by flow cytometry, and the morphology of the cells
`was verified by microscopic examination of Wright-Giemsa­
`stained cytospin preparations. The separated monocytes con­
`tained more than 95% CDT4-positive cells, less than l % CD3-
`positive cells, and less than 1 % CD56-positive cells. Then, the
`
`4 J. H. Zhou. unpublished observa1ions.
`
`monocytes were incubated with 200 units/ml IFN-"( (Genentech,
`San Francisco, CA). 10 ng/ml G-CSF (lmmunex, Seattle, WA),
`or 10 ng/ml GM-CSF (lmmunex) for 24 h before the cytotox­
`icity assay. Monocytes incubated in medium without cytokines
`were included in each assay as the control.
`Flow Cytometric Analysis. FITC- or phycoerythrin­
`conjugated anti-CD3, anti-CD13, and anti-CD56 were pur­
`chased from Becton Dickinson lmmunosystem (San Jose, CA).
`FITC-conjugated mAb 22 (anti-Fc-yRI) was provided by
`Medarax, Inc. (Annandale, NJ). For evaluation of Fc-yRI expres­
`sion, monocytes before and after cytokine incubation were di­
`rectly stained with this panel of rnAbs. For evaluation of im­
`munoconjugate binding, monocytes, as well as cells from HL-
`60, KG- la, and NB4 cell lines, were incubated with 1 µ.g/ml
`BsAb 251 x 22 for I h. The cells were washed twice and
`incubated with FITC-conjugated goat F(ab')z anti-mouse lgG
`(H + L) (Caltag, South San Francisco, CA) for 30 min. All of
`the samples were analyzed by FACScan (Becton, Dickinson).
`FITC- or phycoerythrin-conjugated mouse isotypic immuno­
`globulins were used as controls.
`Assay. The assay was performed in 96-well
`Cytotoxiclty
`round-bottomed microtiter plates (Rainin Instrument Co.,
`Woburn, MA). The target cells were washed once with RPMI
`1640 medium and incubated with 100 µ.Ci sodium [51Cr] chro­
`mate (New England Nuclear, Boston, MA) for I h at 37°C. After
`being washed several times, the cells were resuspended in RPMI
`1640 medium containing 10% FCS to a concentration of 1 x
`105/ml. Monocytes serving
`
`as effector cells were suspended in
`RPMI 1640 medium to a final concentration of I x 107/ml.
`Effector cells (100 µ.I) were added to the first row of cells, and
`a serial dilution was performed with equal volumes of RPMI
`1640 medium. Then, 100 µ.l target cells were added in the wells
`to yield final E:T ratios of 50:1, 25:1, and 12:1. In a standard
`assay, 0.2 µ.g immunoconjugate was finally added to yield a
`final immunoconjugate concentration of I µ.g/ml. SCCL-1, a
`mouse lgG2a mAb directed against transferrin receptor and
`capable of binding to the Fc-yRI through its Fe domain {17),
`was
`included in each assay as a positive control to measure the
`activity of the monocytes. Several other controls with E:T ratios
`of 50: l also were incorporated in each assay, including incuba­
`tion of target cells with the immunoconjugate alone, incubation
`of target and effector cells with no antibody, incubation of target
`and effector cells with the unconjugated parental antibodies, and
`incubation of effector cells with CD33-negative leukemia cells
`in the presence of BsAb 251 x 22. In each assay, a 20-fold
`excess of unconjugated mAb 22 or mAb 251 along with the
`immunoconjugate were incubated together with the target and
`effector cells to determine whether tumor cell lysis could be
`blocked by either of the parental antibodies. Cytotoxicity assays
`also were performed to compare the cytotoxicity mediated by
`the BsAb 251 x 22, mAb 251, and F(ab'h fragment of mAb
`251 at various E:T ratios. To determine whether the presence of
`human immunoglobulin could interfere with the redirected tu­
`mor cell killing, some cytotoxicity assays were performed in
`IOO% human AB serum. After incubation at 37°C for 4 or 18 h,
`the microplates were centrifuged, and the supernatant was col­
`lected for the determination of51Cr release. Maximum lysis was
`achieved by the addition of 100 µ.I 5% NP40 to 100 µ.I target
`cells. The percentage of cell lysis was calculated as: (experi-
`
`Downloaded from
`
`clincancerres.aacrjournals.org
`
`on August 31, 2015. © 1995 American Association for Cancer
`Research.
`
`2 of 8
`
`BI Exhibit 1014
`
`

`

`Clinical Cancer Research 1321
`
`
`
`cells:· __
`Tab/;• I The binding of mAb 22, mAb 251, and BsAb 25 l X 22 to target and effector
`BsAb 251 x 22
`
`mAb 251
`
`mAb 22
`
`M FI
`% Positive M FI
`% Positive
`% Positive
`MFI
`99.6
`151.0
`16.6
`99.5
`HL·6tl
`64.6
`126.4
`56.7
`89. I
`12.9
`54.2
`54.6
`89.2
`NB4
`Negative Negative Negative Negative Negative Negative
`KG-la
`101.3 ----
`79.0
`43.9
`94.0
`27.0
`81.6
`Monocytc
`
`by indirect of cells with mAb 251 (0.5 µg/ml), mAb 3G8 (0.5 µg/ml), or BsAb 251 x 22 (I µg/ml). then
`
`" Flow cytomctric analysis
`staining
`
`
`
`
`in all analyses. were included murine mAb controls with FITC-conjugatcd goal F(ab'), anlimouse lgG (H + L). lsotypc-malchcd
`
`
`
`Tahle 2 Effect of cytokine stimulation on the expression of Fc-yRI
`
`
`
`on monocytes''
`
`too
`
`No. of experiments
`Cytokinc Positivity
`(%) MF!
`i 80
`54 = 3
`No cytokinc 84 = 3
`3
`"' ·;;
`IFN--y
`87 = 9
`3
`150 = 33•
`� 60
`3
`88::: 6 72 = 18
`G-CSF
`70 = 28
`GM-CSF 85 := II
`3
`';; ..
`40
`
`
`
`" Flow cylometric analysis of the expression of FcyRI by direct
`.. c
`mAb 22. Monocytes were
`
`
`staining of monocytes with FITC-conjugated
`e "
`
`
`
`separated from normal peripheral blood and incubated with indicated
`I- 20
`for 24 h before the analysis.
`cytokine
`
`•Significant increase
`
`(P < 0.05) compared to that with no cytokine
`stimulation.
`
`--
`
`-0--
`
`NB4
`
`H[.,.60
`
`O _._.....,..--..---.---..---.-----...---.----.-
`I tO
`0.t
`0 0.00001 O.OOOt 0.001
`0.01
`or BsAb 251 x 22 (µg/ml)
`Concentrations
`
`Fi11. I Cytoloxicity of target
`cells by IFN--y-activated monocytes in the
`
`
`of BsAb 251 X 22. A significant
`
`of various concentrations
`presence
`
`
`increase of target cell lysis was observed with 0.001-IO µg/ml BsAb
`25 I X 22. Bars. SE.
`
`mental cpm - spontaneous release mean cpm) X 100 I (max­
`imum release mean cpm - spontaneous mean cpm). Spontane­
`ous release of s1cr from the target cells was less than 5% of
`maximum release in a 4-h assay and less than 15% in an 18-h
`assay. For dose-response assays. BsAb 251 X 22 was serially
`diluted and added; the E:T ratio in those assays was 50: I .
`Analysis. Each experimental result i n the cy­
`Statistical
`totoxicity assays was obtained in triplicate and reported as the
`mean :t SD. The significance level was determined using the
`paired Student's r test when applicable.
`
`It has been
`Activation of EITector Cells by Cytokines.
`
`well documented that the incubation of monocytes with IFN--y
`increases the expression of Fc-yRI on their surface and stimu­
`lates their cytotoxic activity. Several previous studies indicate
`that G-CSF and GM-CSF also can stimulate the expression of
`RESULTS
`Fc-yRI on circulating monocytes and neutrophils after in vi1•0
`administration. Both of the cytokines have been shown to en­
`Binding of BsAb 251 x 22 to Target and Effector
`Cells.
`The binding of the two parental mAbs (mAb 22 and 251) and
`hance the cytotoxocity of monocytes and neutrophils after in
`BsAb 251 x 22 to target and effector cells was studied by flow
`1·i1m incubation, although the Fc-yRI expression does not in­
`cytometric analysis. HL-60 and NB4 cells were strongly posi­
`crease significantly. We studied the effect of G-CSF and GM­
`tive for CD33 and weakly positive for Fc-yRI. KG-la cells were
`CSF incubation on the expression of FcyRI and cytotoxic ac­
`negative for both CD33 and FcyRI. The BsAb 251 X 22 bound
`tivity of monocytes. Incubation of monocytes with 200 units/ml
`to HL-60, NB4, and monocytes with similar percentages and
`of IFN--y for 24 h resulted in a significant increase of FcyRI
`mean fluorescence intensities. suggesting that rhc process of
`expression, whereas G-CSF and GM-CSF incubation did not
`chemical conjuga tion did not alter the binding characteristics.
`
`increase Fc-y RI significantly (Table 2). W ith prolonged incuba·
`The results of flow cytometric analysis arc presented in Table I.
`tion for 48 h, IFN--y and G-CSF both increased FcyRI expres­
`The BsAb 251 X 22 could be detected on the surface of 90% of
`sion further (MFI = 218 :t 82 and 110 :t 33, respectively),
`HL-60 cells afler 4-h incubation, suggesting that the BsAb
`whereas GM-CSF did not (MFI = 64 :t 16).
`251 x 22 is not readily internalized after binding to CD33
`
`Cytotoxicity of Target Cells in the Presence of BsAb
`molecules on HL-60 cell surface (18). The absence of antigenic
`251 x 22. The concentration dependence of BsAb 251 X 22
`modulation may be important for BsAb 251 x 22 to direct
`required for cytotoxicity was measured. The results of one
`immune effector cells to rarget cells and to induce subsequent
`typical assay are presented in Fig. I. A significant increase in
`cytotoxicity of target cells. The fact that BsAb 251 x 22 does
`tumor cell lysis was observed from 0.00 I µg/ml to IO µg/ml
`not bind to KG-la cells allows us to use these cells as controls
`BsAB 251 X 22. The presence of cytotoxicity at very low
`in the cytotoxicity assay to measure nonspecific tumor cell lysis,
`concentrations of BsAb 251 X 22 may be important for effec­
`tive in 1'ivo
`as well as to validate the specificity of the BsAb-mediated
`immunotherapy, because the concentration of BsAb
`cytotoxicity.
`may be low in tissues.
`
`Downloaded from
`
`clincancerres.aacrjournals.org
`
`on August 31, 2015. © 1995 American Association for Cancer
`Research.
`
`3 of 8
`
`BI Exhibit 1014
`
`

`

`1322 AML Cell Lysis in lhe Presence of BsAb 251 X 22
`
`B.
`HL-60 cells
`
`
`
`A. NB4 cells
`
`B. HL-60 cells
`
`A.
`NB4 cells
`100
`
`�
`80
`-�
`;
`60
`t
`..
`40
`� "
`""
`20
`
`100
`
`80
`
`60
`
`40
`
`20
`
`.•
`.
`. .
`iii''
`
`0
`
`.
`
`-<>-
`
`Donc><I
`
`-
`Dono< 2
`····<>-···
`0o ... 3
`Donor•
`
`--6--
`
`....... Do ... $
`
`• 4h uu.y
`0 18h uuy
`
`too
`
`� 80
`.l!l .,
`� 60
`1i "'
`40
`.. 0
`E
`"
`i;., 20
`
`100
`
`20
`
`~
`� T
`
`12:1 25:1 $0:1
`12:1 25:1 $0:1
`E:T cell ralio
`E:T cell ratio
`of targel cells by IFN--y-activated
`monocytes from
`Fig. 2 Cy101oxici1y
`five donors. The cyloloxicity is dependent on 1he E:T cell ratio in
`time on 1arge1 cell lysis by IFN•'f·activated
`Fig. 3 Effe cl of incubation
`individual donors and varies greatly among different
`donors. Bars, SE.
`
`monocytes from donor 2. A significant increase of target cell lysis was
`
`
`observed in 18-h assays. Bars, SE.
`
`12'1
`
`25:t SO:I
`E:T cell ratio
`
`12:1 25:1 SO:t
`E:T cell ratio
`
`The specific lysis of HL-60 and NB4 cells by IFN--y­
`
`in Table 4, the addition of a 20-fold excess of either parental
`
`
`
`
`
`monocytes is summarized in Fig. 2. The cytotoxicity
`activated
`
`
`antibody significantly inhibited the increment of cytotoxicity
`
`
`
`time. At an E:T and incubation was dependent on E:T ratios
`
`
`
`mediated by BsAb 251 x 22. As expected, the presence of
`
`ratio of 50:1, more than 70% of target cells were lysed in the
`
`
`human AB serum did not interfere with the cytotoxicity medi­
`
`
`of BsAb 251 x 22. Variations of cytotoxicity among
`presence
`
`
`ated by BsAb 251 X 22 because the binding of mAb 22 to
`
`
`
`donors and differences in susceptibility of target cells
`individual
`
`
`Fc-yRI occurs outside the Fe binding site. The cytotoxicity
`
`
`to lysis were observed. As shown in Fig. 2, monocytes from
`
`
`
`mediated by mAb SCCL-1 (an IgG2a mAb) was blocked sig­
`
`
`toward both HL-60 and donor 5 had the highest cytotoxicity
`nificantly
`
`
`by human AB serum, presumably because the pres­
`NB4 cells, whereas monocytes from donor 3 had a much higher
`
`
`ence of serum immunoglobulin could compete with the Fe
`
`cy1otoxicity toward NB4 cells (40-75%)
`than did HL-60 cells
`to FqRI (data
`
`portion of the IgG2a SCCL-1 mAb for binding
`
`
`The expression of FqRI on monocytes after IFN--y
`(25-45%).
`not shown).
`
`
`
`An additional control compared the effects of the
`
`
`
`
`incubation was comparable in all cytotoxicity assays, and the
`
`mAb 251 and its F(ab'h fragments to BsAb 251 x 22
`parental
`
`BsAb 251 X 22 used in these assays was derived from one
`
`
`
`on the cytotoxicity of leukemia target cells (Fig. 4). The cyto­
`
`
`
`conjugation. The effect of incubation time on the cytotoxicity
`
`
`
`toxicity mediated by mAb 251 and its F(ab'h fragments was
`
`
`
`assay was important. A comparison of cytotoxicity from a 4-h
`
`
`comparable wi1h that mediated by monocytes alone, whereas
`
`
`and an 18-h assay is presented in Fig. 3. In these assays, the
`
`
`
`of BsAb 251 x 22 significantly enhanced the cytotoxicity
`
`background leak of 51Cr was less than 5% of the maximal
`
`leukemia targets.
`
`release in the 4-h assay and less than 15% of the maximal
`
`
`
`release in the 18-h assay. The percentage of tumor cell lysis was
`DISCUSSION
`
`
`
`significantly increased after an 18-h incubation. This observa­
`Although
`the treatment of AML has been improved signif­
`
`
`
`tion, combined with the fact that a higher E:T ratio is required
`
`
`
`icantly by various new chemotherapeutic agents and regimens,
`
`
`
`
`for cytotoxicity, suggests that the production of cytotoxic cyto­
`
`
`
`
`
`relapse as a result of the persistance of minimal residual disease
`
`
`kines may be involved in target cell lysis.
`
`
`
`after chemotherapy remains a major problem for these patients.
`The Effect or Cytoltine Activation on Cytotoxicity of
`
`
`
`
`
`
`Efforts such as allogeneic bone marrow 1ransplantation have
`
`
`
`A series of assays was performed to compare
`Target Cells.
`
`
`been made to further improve the outcomes of the current
`
`
`
`
`the effects of cytokine incubation on the enhancement of cyto­
`
`
`
`therapy. In addition, autologous bone marrow transplantation
`
`
`
`
`
`toxicity of target cells. In general, the cytotoxicity of HL-60 and
`
`
`
`with AML. has been used as consolidation therapy for patients
`
`NB4 cells by monocytes incubated with G-CSF and GM-CSF
`
`
`
`Gene-marking experiments of bone marrow 1ransplants in
`
`
`
`was comparable to that mediated by monocytes incubated
`
`
`AML have suggested that residual leukemic cells in the
`
`
`
`
`with IFN--y. All three cytokines significantly increased the
`
`
`
`
`transplant can contribute to the relapse of leukemia ( 18).
`
`
`
`cytotoxic potency of monocytes. The results are summarized
`
`
`
`
`Various purging techniques have been developed, including
`in Table 3.
`
`
`treatment of bone marrow with cytotoxic drugs, mAb plus
`Specificity
`BsAb 251 x 22.
`or Cytotoxicity Mediated by
`
`
`
`
`complement, and immunomagnetic bead separation (19). All
`
`
`To determine whether target cell lysis was induced specifically
`
`
`of these strategies have met with some measure of success.
`KG-la cells by the presence of BsAb 251 x 22, CD33-negative
`
`
`
`
`
`
`However, considerable numbers of patients still relapse be­
`
`
`
`assays. Lysis of were used as target cells in some cytotoxicity
`
`
`
`
`cause of either failure of the primary therapy to eliminate all
`
`KG-la cells in the presence of BsAb 251 X 22 was very low,
`
`
`
`in vivo or failure to purge all malignant cells from the
`disease
`
`from 3.7% :t 1.2% (without IFN--y incubation) to
`ranging
`autograft.
`
`
`
`7.7% :t 1.0% (with IFN--y incubation). The specificity of target
`In this study, we examined an alternative strategy to elim­
`
`
`
`
`
`cell lysis can be shown further by blocking the cytotoxicity with
`
`
`inate residual leukemia cells by directing host immune effector
`
`
`
`
`excess concentrations of either parental antibody. As illustrated
`
`Downloaded from
`
`clincancerres.aacrjournals.org
`
`on August 31, 2015. © 1995 American Association for Cancer
`Research.
`
`4 of 8
`
`BI Exhibit 1014
`
`

`

`
`
`in the presence of BsAb 251 x 22"
`
`cytotoxicity Table J Effect or cytokine incubation on monocyte-mediatcd
`Cytokine
`
`Clinical Cancer Research 1323
`
`E:T
`NB4 cells
`12:1
`
`25:1
`
`50:1
`
`HL-60 cells
`12:1
`
`Donor
`
`IFN-'I
`
`G-CSF
`
`GM-CSF
`
`No cytokinc
`
`3
`4
`5
`3
`4
`5
`3
`4
`5
`
`40.9 :!: 2.4
`19.6 :!: 0.4
`34.9 :!: 4.3
`65.8 :t 3.4
`33.3 :t 0.6
`56.0 :!: 2.8
`69.0 :t 1.3
`44.5 :!: 4.8
`70.5 :t 4.1
`
`43.6 :!: I. I
`24.2 :!: I.I
`32.2 :!: 5.5
`55.0 :!: 3.4
`32.7 :t 1 . 7
`46.4 :!: 2.1
`63.5 :!: 0.7
`45.8 :!: 3.5
`57.8 :!: 2.5
`
`41.5 :!: 2.0
`15.5 :!: 0.5
`ND'
`56.5 :!: I. 7
`27.3 :!: 0.7
`No•
`61.7 :!: 2.6
`40.4 :!: 4.4
`No•
`
`5.9 :!: 0.5
`3.3 :!: 0.9
`7.1 :t 1.8
`6.1 :t 1.5
`5.5 :t 0.4
`10.2 :t I.I
`6.5 :!: 1.2
`5.3 :t I.I
`13.9 :!: 1.3
`
`25:1
`
`50:1
`
`3
`4.4 :!: 0.5
`26.2 :t 0.6
`29.4 :!: 1.8
`26.5 :!: 0.9
`35.4 :!: 1.3
`4
`29.I :!: I.I
`28.2 :!: 1.0
`3.6 :!: 0.9
`5
`60.7 :!: 4.1
`48.6 :!: 4.9
`53.9 :!: 4.4
`0.6 :!: 0.1
`3
`5.8 :!: 0.7
`41.4 :!: 1.6
`37.3 :!: 0.7
`35.6 :!: 2.0
`41.5 :!: 1.7
`5.2 :t 0.5
`33.0 :!: 2.1
`34.0 :t 1.2
`4
`61.5 :!: 4.4
`74.5 :!: 5.1
`59 4 :t 2.5
`5
`2.1 :!: 0.2
`49.6 :t 1.4
`46.0 :t 0.7
`51.3 :!: 0.5
`3
`6.3 :!: 0.4
`5.0 :!: 0.8
`36.7 :!: 2.1
`4
`42.9 :t 1.0
`52.0 :!: 1.8
`5
`75.9 :t 1.3
`10.2 :!: 0.2
`82.7 :!: 3.5
`78.0 :t 6.7
`"Monocytes were separated from three different donors. Percentage of target cell lysis is expressed as mean :t SD of triplicates.
`
`•ND. not done.
`
`by its Table 4 Inhibition or BsAb 251 x 22-mediated cytotoxicity
`
`
`
`
`parental antibody"
`
`Inhibition (%)
`
`Experiment
`
`has been humanized as well.5 Therefore, the construction of a
`humanized BsAb for immunotherapy targeting CD33 and CD64
`is feasible.
`Activation of immune effector cells by cytokines was cru­
`
`mAb 251
`mAb 22
`cial for the cytotoxicity of target cells. Monocytes without
`90
`95
`preincubation with cytokfoes had much less cytotoxicity than
`96
`86
`did those with preincubation. Although
`IFN--y incubation
`"The conce ntration of BsAb 251 x 22 was I µg/ml. and the
`yielded the highest expression of Fc-yRI, its application in vivo
`
`concentration of each parental mAb was 10 µg/ml. The E:T ratio in
`for this purpose has not been studied extensively. In this study,
`
`was 50: I. The cytotoxicity
`mediated by BsAb 251 X
`these experiments
`
`22 was defined as 100%. and the cytotoxicity observed by mixing target
`we compared the effect of IFN--y, G-CSF, and GM-CSF incu­
`
`
`cell s with lFN-'f·activated monocytes without any antibody was defined
`bation on the expression of Fc-yRI and cytotoxic activity of
`as 0%. Th e percentage or inhibition
`
`was calculated accordingly.
`monocytes. After 24-h incubation, all three cytokines signifi­
`cantly enhanced the cytotoxicity of monocytes in the presence
`of BsAb 251 X 22. The cytotoxicity of target cells by mono­
`cytes incubated with G-CSF and GM-CSF is comparable that of
`these cells incubated with IFN--y. The expression of Fc-yRI was
`on monocytes incubated with IFN--y for
`significantly increased
`24 h, but not on monocytes incubated with G-CSF or GM-CSF
`for 24 h. However,
`the expression of Fc-yRI was increased
`significantly on monocytes incubated with G-CSF for 48 h. It
`has been previously shown that incubation of monocytes with
`G-CSF or GM-CSF in vitro could increase their cytotoxic ac­
`tivity (5, 7, 9). and recent data showed that the expression of
`be increased after infusion of G-CSF in vivo (9). It
`Fc-yRI could
`has also been reported that IFN--y-activated monocytes can be
`used for in vitro purging of clonogenic leukemic cells (23).
`These observations are important for clinical
`application of our
`immunotherapeutic approach, because both G-CSF and GM­
`CSF have been widely used after chemotherapy and/or bone
`
`cells toward leukemia cells
`
`via a BsAb. Unconjugated mouse
`mAbs alone have been studied for their antitumor effects in
`patients with AML and other malignancies by ADCC and other
`effector mechanisms (20, 21). Unconjugated mAbs suffer from
`several disadvantages: one is that the binding of a mAb to Fe
`receptors on effector cells may be competed by circulating
`immunoglobulin. This nonspecific binding of circ ulating immu­
`noglobulin
`to Fc-yRI on immune effector cells can block the
`ADCC triggered by unconjugated
`antibody. This problem can
`be overcome by BsAb 251 X 22 because mAb 22 binds with
`high affinity to Fc-yRI outside the binding site of Fe domain of
`immunoglobulins (22). BsAb 251 x 22 can bind and activate
`Fc-yRI without the interference of the high concentration of
`circulating immunoglobulin. The other major disadvantage of
`mouse mAbs is that they may induce the formation of human
`antimouse antibody, thereby abolishing the therapeutic effect of
`the mAb. The development of recombinant humanized versions
`of mouse mAbs can overcome this problem. A humanized
`version of anti-CD33 mAb has been reported ( 14 ), and mAb 22
`
`5 R. Graziano. personal communication.
`
`Downloaded from
`
`clincancerres.aacrjournals.org
`
`on August 31, 2015. © 1995 American Association for Cancer
`Research.
`
`5 of 8
`
`BI Exhibit 1014
`
`

`

`1324 AML Cell Lysis in the Presence of BsAb 251 X 22
`
`·;;; ~ � 60
`60 �
`
`:; "'
`
`40
`"'
`c
`El
`:I 20
`!-
`
`Q....-�
`
`40
`
`20
`
`0
`
`0
`
`25:1 50:1
`12:1
`E:T cell ratio
`
`50:1
`25:1
`12:1
`E:T cell ratio
`
`Fig. 4 Comparison of the cytotoxicity mediated by BsAb 251 X 22,
`mAb 251, F(ab'h fragment of mAb 251, and controls. Monocytes were
`separated from a single donor and incubated with IFN·'I· In controls
`(£>), IFN-'l·activated monocytes were mixed with target cells without
`any antibody. The concentrations of BsAb 25 I X 22 (•), mAb 251 (0),
`and F(ab')z fragment of mAb 251 (0) were I µg/ml. The presence of
`mAb 251 and F(ab'h fragment of mAb 251 did not increase the
`cytotoxicity compared with that of controls. The addition of BsAb
`251 x 22 increased the cytotoxicity significantly. One of three such
`experiments with similar results is shown. Bars, SE.
`
`marrow transplantation to stimulate the recovery of myeloid
`hematopoiesis.
`On the other hand, the expression of the CD33 antigen is
`restricted to myeloid leukemia, monocytes, and myeloid pro­
`genitor cells. A recent clinical trial in AML patients using
`1311-conjugated to a humanized version of anti-CD33 mAb
`showed rapid localization to target cells and efficient cytoreduc­
`tion of leukemic blasts with little systemic toxicity (14). Anti·
`CD33 mAb covalently linked to ricin toxin or in combination
`with rabbit complement and chemotherapy has been used to
`purge bone marrow ex vivo (24-26). All of these findings
`suggest that the CD33 may be an excellent molecule for targeted
`immunotherapy. Of note, the expression of CD33 on the mono·
`cyte effector cells (albeit with lower cell surface density than on
`AML cells) in the experiments we report herein did not affect
`their ability to target CD33-positive leukemia cells. In other
`words, the possibility that the CD33 molecules on monocytes
`could act as targets for their own cytotoxicity did not appear to
`be an obstacle. Whether there was any ADCC mediated by
`monocytes against monocytes remains unknown: we intend to
`study this possibility in the future. Although the maximal killing
`of target cells required higher E:T ratios, significant tumor cell
`lysis was observed at lower E:T ratios. Because the cytotoxicity
`in an ADCC assay is measured by choosing a window of time
`in the ongoing cytotoxic process, it probably does not represent
`the full potential for in vivo cytotoxic activity. The effector cells
`may work more efficiently in vivo, when they are in their natural
`states, can interact with other immune effector cells, and have
`much more time to contact target cells and mediate cytotoxicity.
`There are several potential applications for BsAb 251 X 22
`in the treatment of AML. One application is as an adjuvant
`
`immunotherapy to eliminate minimal residual leukemia cells.
`Although the majority of AML patients will enter a clinical
`remission after standard chemotherapy, a significant number of
`patients will relapse eventually. The administration of BsAb
`251 x 22, with the combination of G-CSF or GM-CSF, will
`enhance the recovery of hematopoiesis and, at the same time,
`mobilize and direct immune effector cells to eliminate the
`residual leukemic cells. BsAb 251 x 22 also may play a role in
`the setting of allogeneic bone marrow transplantation by aug­
`menting the graft-versus-leukemia effect. Finally, it can be
`applied in