Monoclonal Antibodies to Myeloid Differentiation Antigens: In Vivo
`Studies of Three Patients With Acute Myelogenous Leukemia
`
`By Edward D. Ball, George M. Bernier, Gibbons G. Cornwell Ill, 0. Ross McIntyre, Joseph F. O'Donnell, and
`Michael W. Fanger
`
`Three patients with acute myelogenous leukemia (AML) in
`relapse were treated with intravenous infusions of one or
`more purified murine monoclonal antibodies (MoAbs) spe—
`cific for differentiation antigens on normal and malignant
`myeloid cells. Three of the MoAbs used were lgM immuno-
`globulins that react with glycolipids. while the fourth, an
`IgGZb. reacts with a protein antigen. Peripheral blood
`leukemia cell counts decreased significantly, but transient-
`ly, during treatment. Evidence of in vivo binding of each
`MoAb to leukemia cells was obtained. although two of the
`four MoAbs could not be detected in the plasma following
`infusion, perhaps due to circulating blocking factors. Anti—
`genic modulation was not encountered in these studies.
`
`However, the induction of human antibody to murine MoAb
`was observed in one patient who was treated over a
`70-day period. Toxicities encountered were minimal and
`included fever (3 patients). back pain (1 patient). and
`arthralgias and myalgias (1 patient). This is the first
`reported clinical trial of (1) lgM MoAbs, (2) MoAb therapy
`in patients with AML. (3) combinations of MoAbs directed
`toward different myeloid differentiation antigens. and (4)
`MoAbs directed to glycolipids. The relative lack of toxicity
`and the positive effects of MoAb treatment in the reduc-
`tion of leukemia cell counts permit the continued study of
`more innovative approaches to the treatment of AML with
`MoAbs.
`
`targeted to
`first report of human MoAb treatment
`glycolipids, and the first
`treatment of AML with
`MoAb directed to myeloid difierentiation antigens.
`The lack of serious toxicity with these reagents permits
`numerous
`future applications of
`these and other
`MoAbs.
`
`MATERIALS AND METHODS
`
`Patient Selection
`
`Patients with AML who were refractory to standard chemother-
`apy or who were poor risks for further chemotherapy were consid-
`ered for MoAb treatment. Treatment of patients with MoAb was
`approved by the Committee for the Protection of Human Subjects,
`and all patients gave informed consent
`for all aspects of the
`therapeutic protocol.
`
`Case Reports
`Patient 1. R3. was a 27-yr-old white female who developed
`AML in September I981, At that time. her WBC was 8.600/u1 with
`2% blasts, and her bone marrow was hypercellular. with 15% blasts
`and 8% promyelocytes. Auer
`rods were present, and she was
`classified as FAB type M2.'° She achieved a complete remission
`(CR) with cytosine arabinoside (Ara-C), daunorubicin (DNR), and
`6-thioguanine (6-TG), but relapsed in February I982. A second
`
`From the Departments of Medicine and Microbiology, Dart-
`mouth—Hitchcock Medical Center, and the Norris Cotton Cancer
`Center, Hanover, NH.
`Supported in part by Grants CA 31918, CA 23l08. CA 31888.
`and A1 [9053 awarded by the National Cancer Institute and the
`Institute of Allergy and Infectious Diseases. DHHS, respectively.
`and by the York Cross of Honour Research Foundation. The
`cytofiuorograph was the generous gift of the Fannie E. Rippel
`Foundation and is partially supported by the core grant of the
`Norris Cotton Cancer Center (CA 23108).
`Submitted April 25, I983; accepted July 9, 1983.
`Address reprint requests to Dr. Edward D. Ball, Department of
`Microbiology. Dartmouth Medical School. Hanover, NH 03756.
`© 1983 by Grune & Stratton. Inc.
`0006—497 [/83/6 206-0009$Ol .00/0
`
`
`
`EVERAL lNVESTlGATORS have reported in
`vivo treatment of human leukemias and lympho-
`mas with murine monoclonal antibodies (MoAbs)
`directed to a variety of tumor-associated antigens."5
`Patients with acute lymphocytic leukemia (ALL),l
`chronic lymphocytic leukemia (CLL),2 T-cell lympho-
`mas,3 and nodular lymphocytic lymphoma“'5 have all
`been treated in this manner, with variable results.
`Several obstacles to such therapy have been identified,
`including temporary internalization (modulation) of
`the cell surface antigen targeted for therapy"6 and
`circulating blocking factor(s).2‘5 While most trials of
`MoAb therapy have produced only temporary
`decreases in tumor cell counts, toxicity has been mini-
`mal. Of interest is the report of the successful induc-
`tion of complete remission in a patient with a B-cell
`nodular lymphocytic lymphoma using a monoclonal
`antibody directed to the idiotype expressed on the
`patient’s tumor cells.5
`We have recently prepared four MoAbs reactive
`with different normal myeloid differentiation antigens
`that are also expressed on leukemia cells from patients
`with acute myelogenous leukemia (AML).7'9 In this
`report, we describe the treatment of three patients with
`acute myelocytic leukemia (AML) with these MoAbs.
`These studies were conducted to determine (1) the
`safety of MoAb administration in patients with AML,
`(2) changes in concentrations of circulating blasts and
`polymorphonuclear cells (PMNs), (3) the presence of
`free and cell-bound MoAb after therapy, (4) rate of
`clearance of circulating MoAb, and (5) the presence of
`factor(s) in patient and normal plasma that block
`MoAb binding to cells. Three of
`the antibodies
`employed in this study are lgM and react with glyco-
`lipids. To our knowledge, this article describes the first
`trial of in vivo MoAb treatment with lgM MoAbs, the
`
`Blood, Vol, 62, No. 6 (December), 1983: pp. 1203—1210
`
`Celltrion, |nc., Exhibit 100
`
`1 of 9
`
`Celltrion, Inc., Exhibit 1009
`
`

`

`1204
`
`BALL ET AL.
`
`remission was attained with intermediate doses of Ara-C and
`L-asparaginase,“ but the patient relapsed and MoAb therapy was
`started.
`is a 65-yr-old white male who presented in
`Patient 2. PS.
`February 198] with fatigue, sore throat, and a platelet count of
`l2,000/ul. Bone marrow was diagnostic ofAML (FAB M l ). and his
`leukemic cells exhibited trisomy No. 8 on chromosomal analysis. He
`achieved a CR with Ara-C, DNR, and 6-TG, but relapsed in
`January 1982. Following reinduction with intermediate dose Ara-C,
`transient aplasia complicated by bacterial and fungal
`sepsis
`occurred, but
`leukemic cells reappeared.
`In May I982, he was
`treated with cis-retinoic acid (provided by Hoflmann-LaRoche, lnc.,
`Nutley, NJ) in the hope of inducing differentiation. He developed
`significant cutaneous toxicity, no improvement
`in his peripheral
`counts, and increasing transfusion dependency. He was begun on
`MoAb therapy in September 1982.
`Patient 3. LS. was a 65-yr-old white male who was well until
`September 1981, when he developed fever, adenopathy, and a rash.
`At that time, he was found to have a WBC of 46,000/ul, with 48%
`blasts and a bone marrow consistent with acute myelomonocytic
`(FAB M4) leukemia. A CR was attained with Ara-C and DNR, and
`he was maintained with Ara-C, DNR, 6-TG, and vincristine. In
`June I982, he had a biopsy-proven recurrence ofleukemia in the skin
`of his right buttock and posterior thigh. At that time, his peripheral
`blood and bone marrow were normal, and he responded to 1,500 rad
`of radiation therapy. In October I982, he developed diffuse arthral-
`gias and increasing cutaneous nodules of the face and neck, which
`biopsy showed to be leukemic. His bone marrow remained normal,
`and he was treated with MoAbs.
`
`Monoclonal Antibodies
`
`The MoAbs used in this study included PMN 6, PMN 29, and
`PM-Sl. all IgM class immunoglobulins, and AML-2-23, an IgGZb
`MoAb (Table l). PMN 6 and PMN 29 react with mature granulo-
`cytes, while PM-8l and AM L-2-23 react with both granulocytes and
`monocytes. Each of these MoAbs are cytotoxic in the presence of
`rabbit serum to normal and leukemic myeloid cells.” None of these
`four MoAbs are reactive with the colony-forming unit—granulocyte/
`monocyte (CFU-GM) or burst-forming unit—erythroid (BFU-E).9
`PMN 6 and PMN 29 react with blast cells from patients with AML
`of the M4 subtype, PM-SI with all subtypes of AML, and AML-
`2-23 with the M4 and M5 subtypes.” An IgM and an IgGZb MoAb
`of irrelevant specificity were used as controls for in vitro studies.
`
`Antibody Purification
`Hybridoma cells producing each of the above antibodies were
`grown in the peritoneal cavity of pristane-primcd BALB/c mice.
`Ascitic fluid was collected aseptically and pooled. IgM MoAbs were
`purified by gel filtration through Sepharose CL-6B (Pharmacia Fine
`Chemicals, Piscataway, NJ) using sterile pyrogen-free phosphate-
`bufl'ered saline (PBS) (pH 7.4) as running bufler. AML-2-23 was
`purified on a protein-A—Sepharose column by elution at pH 3.5.”
`All antibody preparations were passed through a 0.2-u filter and
`stored at 4°C. Cultures for aerobic and anaerobic bacteria and fungi
`
`Table 1 . Monoclonal Antibodies Employed in This Study
`Specificity
`
`MoAb
`
`PM-8‘l
`AML—2-23
`PMN 29
`PMN 6
`
`Isotype
`
`IgM
`IgGZb
`IgM
`IgM
`
`
`
`were negative. Endotoxin assays, performed by the Limulus amoeba-
`cyte lysate assay (Associates of Cape Cod, lnc., Woods Hole, MA),”
`consistently revealed less than 2 ng endotoxin/mg MoAb.
`
`Cell Separations
`Leukemia cells from patients treated in this study and cells from
`normal volunteers were harvested from the peripheral blood as
`previously described.” Leukemia cells were separated by Ficoll-
`Hypaque gradient centrifugation. Normal polymorphonuclear leu-
`kocytes
`(PMNs) were harvested by sedimentation of Ficoll-
`Hypaque dense cells in 2% dextran.
`
`Antibody Binding Assays
`Binding of MoAbs to leukemic and normal cells was determined
`by indirect immunofluorescence and flow cytometry. Target cells
`(2 x 10") were incubated with 50 ul of purified MoAb at 20 ug/ml
`for 30 min at 4°C,
`followed by the addition of a fluorescein
`isothiocyanate (FITC) coupled goat F(ab’)1 antibody directed to
`mouse immunoglobulin (FITC-GAM)
`(Boehringer-Mannheim,
`Indianapolis, IN) for another 30 min at 4°C. The cells were then
`analyzed for fluorescence on the Ortho (Westwood, MA) Cytofluor-
`ograph System 50H, with multichannel distribution analyzer 2103
`and Ortho 2150 computer systems. Positive binding was defined as
`fluorescence on MoAb-reacted cells greater than on control MoAb-
`reacted cells.
`
`Blocking Studies
`Plasma from leukemic patients and normal controls was studied
`for the presence of blocking factors that might interfere with the
`binding of MoAb to cell surfaces. Plasma was incubated with equal
`volumes of varying concentrations of purified MoAb for 15 min at
`4°C prior to addition of this mixture to target cells. Following
`incubation at 4°C for 30 min. FITC-GAM was added for an
`additional 30 min at 4°C. Cells treated in this manner were then
`analyzed by flow cytometry.
`
`Determination of Patient Antibody Production to
`Mouse Antibody by Enzyme-Linked lmmunosorbent
`Assay (ELISA)
`
`MoAbs PM-8 l , PMN 29, and AML-2-23 were applied to individ-
`ual wells of 96-well fiat-bottom plates at 20 ug/ml in PBS and the
`plates incubated at room temperature (RT) overnight. The plates
`were washed with PBS and incubated with 5% bovine serum albumin
`(BSA) in PBS for 1 hr at 37°C. Control wells were treated with 5%
`BSA only. After washing, serially diluted normal and patient plasma
`(pre-, intra-, and posttreatment) were applied in duplicate to MoAb-
`treated and control wells. The plates were incubated for 2 hr at 37°C
`and washed with PBS. Alkaline phosphatase-labeled affinity-
`purified goat anti-human antibodies specific for either IgM or IgG
`(Kirkegaard & Perry Labs, lnc., Gaithersburg, MD) were added
`and the plates incubated at RT for 16 hr. The plates were again
`washed and p-nitrophenyl phosphate disodium (Sigma Chemical
`Corp., St. Louis, MO) was applied. The plates were observed for the
`appearance of yellow color, and the optical density (OD) at 405 nm
`of each well was determined on an ELISA reader after l5 min. The
`mean OD of duplicates of each serum dilution assayed against BSA
`alone (background) was subtracted from that of the same dilution
`assayed against MoAb.
`
`Monoclonal Antibody Administration
`Purified MoAb was administered by intravenous infusion. The
`dose to be administered was diluted into a final volume of 500 ml of
`
`Celltrion, lnc., Exhibit 100
`
`Normal Cells
`
`AML
`Subclass
`
`Reference
`
`M 1-M5
`Granulocyte/monocyte
`Monocyte/granulocyte M4-M 5
`Granulocyte
`M4
`Granulocyte
`M4
`
`8
`7 ,9
`7,9
`7 ,9
`
`2 of 9
`
`Celltrion, Inc., Exhibit 1009
`
`

`

`MONOCLONAL ANTIBODY THERAPY OF AML
`
`1205
`
`Table 2. Binding of MoAb to Leukemia Cells From AML Patients'
`
`Monoclonal Antibody (96 Cells Positive)
`PMN 29
`PM-8‘l
`PMN 6
`AML-2—23
`
`,
`Patient
`(FAB Class)
`
`1(M2)
`2(M 1)
`3(M4)
`
`oton
`11 (7)
`54(110)
`
`0(0)
`59 (61)
`70(110)
`
`86 (41)
`75 (99)
`92(120)
`
`0(0)
`22 (20)
`94(135)
`
`'Binding was determined by indirect immunofluorescence and flow
`cytometry. MoAb—treated cells were considered positive when fluores-
`cence exceeded the background level detected on cells treated with
`control MoAb.
`
`fThe number in parentheses is the mean fluorescence intensity (MFI)
`of MoAb-treated cells corrected for the MFI of control MoAb-treated
`cells.
`
`Blast cells from patient 3 obtained at his first relapse
`reacted with all four of the MoAbs. Binding studies of
`PM-8l to normal PMNs and blast cells revealed that
`
`respectively)
`both cell populations (2 x 10" cells,
`showed saturable binding at approximately 3 ug/ml of
`the PM-8l MoAb.
`
`Effects of Treatment
`
`Patient 1
`
`Following the first infusion (60 mg of PM-8l over 8
`hr),
`the peripheral blood blast cell count fell from
`54,000/ul to 22,000/ul (Fig. 1). Twelve hours later,
`however, the blast cell count returned to 48,000/al.
`The next administration of PM-Sl (60 mg) produced a
`depression in the blast count
`to l9,000/p.l. Again,
`however, the blast count rose to 43,000/ul within 12 hr
`and to 55,000/pl 24 hr later. While more antibody was
`being prepared, the patient was treated with DNR (25
`mg/sq m), which decreased the blast count to 5,000/ul
`before it began to rise. On day 10, a 30-mg dose of
`I’M—81 was administered, and her blast count fell from
`
`
`
`normal saline containing 5% human albumin. Following preparation
`of the final solution to be infused into the patient. culture for bacteria
`and fungi was performed. Solutions ofantibody were usually admin-
`istered over 8 hr, while the patient was being monitored closely for
`changes in vital signs or symptoms of allergic reaction. The duration
`of infusion in patient 2 was 24 hr on one occasion and 7 days on
`another (see Fig. 3). In order to minimize potential allergic reac-
`tions. patients 2 and 3 were premedicated with 50 mg diphenhy-
`dramine hydrochloride and l00 mg hydrocortisone immediately
`before each MoAb infusion.
`
`Rationale for MoAb Dose and Infusion Rate
`The dosage of MoAb was initially calculated based on the amount
`of MoAb that saturated a given number of leukemia cells in vitro
`and an estimate of the patient's leukemia cell burden. We infused
`each dose over 8 hr because other investigators had noted less
`toxicity with slow compared to rapid infusions.I We modified this
`approach in patient 2 after the first series of infusions revealed
`favorable effects on the blast cell counts, with little toxicity. There-
`fore. we increased both the amount and the duration of the MoAb
`infusion in an effort to produce a more sustained depression in the
`blast count.
`
`Study Parameters
`Patients treated with MoAbs were observed for changes in the
`white blood cell and differential counts, hemoglobin, platelet count.
`prothrombin time. partial
`thromboplastin time. urinalysis, serum
`creatinine, alkaline phosphatase. glutamic oxaloacetic transaminase
`(GOT). glutamic-pyruvic transaminase (GPT), lactic dehydrogen-
`ase (LDH). and isocnzyme profile and 5' nucleotidase.
`Free MoAb in the serum of patients during and after treatment
`was measured by incubating PMNs sequentially with plasma sam-
`ples and FlTC-GAM before, during. and after therapy. and studied
`by flow cytometry. Blast cells from patients treated with MoAb were
`harvested at various times during treatment and studied for the
`presence of cell surface mouse antibody by adding FlTC-GAM.
`Aliquots of the same cell preparations were treated in parallel, with
`the addition of purified MoAb followed by the FlTC-GAM and flow
`cytometric analysis to quantify free binding sites for MoAb.
`A skin nodule biopsy was performed on patient 3 (who had
`leukemia cutis) during the infusion ofthe AML-2-23 antibody. Cells
`from the skin nodule were dissociated by forceps and passed through
`a steel screen to make a single—cell suspension. Aliquots of these cells
`were then incubated either with FlTC-GAM or with purified
`AML-2-23 antibody followed by FlTC-GAM and studied by flow
`cytometry.
`
`RESULTS
`
`Monoclonal Antibody Binding
`
`The results of pretreatment cytofiuorographic anal-
`ysis of MoAb binding to the leukemia cells from the
`patients treated in this study are shown in Table 2. The
`leukemia cells of patient I reacted strongly with the
`PM-8l MoAb but not with PMN 6, PMN 29, or
`
`AML-2-23. The cells of patient 2 reacted strongly with
`PM-81 and, to a lesser extent, with PMN 29. At the
`
`time of this analysis, significant numbers of more
`differentiated cells (promyelocytes and myelocytes)
`were present, which probably accounted for the rela-
`tively strong reaction of PMN-29 with these cells.
`
`88
`CELLcouumux10' 8
`
`
`
`\—-.
`f_________
`11
`12
`
`5
`
`4
`DAYS
`
`Fig. 1. Circulating blast count in patient 1 following MoAb
`I’M-81 infusion. The peripheral blood blast cell count is plotted
`against time in relation to infusion of I’M-81 MoAb (I). The
`amounts of I’M-81 infused were 60. 60. 30. and 20 mg. respec—
`tively. during each of the indicated treatments. A dose of DNR (85
`mg) was given on day 5. The blast cell count decreased to 9.000/ul
`by day 8 and remained at this level for the 3 days prior to the next
`MoAb infusion.
`
`Celltrion, Inc, Exhibit 100
`
`3 of 9
`
`Celltrion, Inc., Exhibit 1009
`
`

`

`BALL ET AL.
`
`Blast cells harvested at 1 hr and 5 hr during the first
`MoAb infusion and 1 hr after completion of the
`infusion revealed detectable amounts of the PM-81
`
`Patient 2
`
`The first treatment involved a series of infusions of
`
`PMN 6, PMN 29, and PM-8l MoAbs; the second
`treatment consisted of PMN 29 and PM-81; while the
`final treatment involved the use of PM-81 only (Fig.
`3). At the onset of therapy, the patient had a white
`blood cell count of 18,000/ul, with 22% blasts and 50%
`myeloid cells at various levels of maturation. The
`treatment plan was to use two different MoAbs (PMN
`6 and PMN 29) to remove more mature myeloid cells
`before using PM-81, which reacted more strongly with
`the patient’s blast cells. Infusion of PMN 6 resulted in
`a drop of the mature myeloid cell (PMNs, bands,
`myelocytes, and promyelocytes) count from 9,000/pl
`
`..
`to— con
`
`: (
`
`Id
`
`fit
`
`
`
`CELLNUMBER
`
`/ in viva BINDING
`
`a / In mmBINDING
`
`FLUORESCENCE
`INTENSITV
`
`Fig. 2. Binding of MoAb PM-81 to blast cells from patient 1
`obtained during PM-Bl
`infusion. Blast cells were separated by
`FicolI-Hypaque gradient centrifugetion and incubated either with
`FITC-GAM alone (in vivo binding) or with saturating amounts of
`PM-81 followed by FlTC-GAM (in vitro binding) and analyzed on
`the cytofluorograph. Cells (CON) obtained before therapy were
`incubated with control IgM MoAb and FITC-GAM and compared to
`cells reacted with I’M-81 MoAb. Evidence of in vivo binding is seen
`from this analysis. although saturation of all PM-81 binding sites
`was not achieved. The degree of binding of control cells incubated
`with I’M-81 (not shown) is identical to that of cells exposed in vivo
`to PM-81 . and these in vitro binding curves are superimposable.
`
`9,000/ul to 5.000/ul. Another dose of PM—Sl (20 mg)
`produced a modest decrease in white count to 4,000/
`ul. Due to the patient’s grave clinical condition, which
`included probable sepsis, the patient and her physi-
`cians elected to withhold any further therapy, and she
`expired 1 wk later.
`During the first infusion of the PM-8l antibody to
`patient 1, she developed an urticaria] eruption on her
`legs, which responded promptly to a single dose of
`diphenhydramine hydrochloride (50 mg iv) and
`hydrocortisone (100 mg i.v.). Prior to the next three
`infusions, she was premedicated with these drugs and
`experienced no further skin reactions or other adverse
`effects, with the exception of mild febrile reactions.
`
`20 Dm-
`l: = PMN 6
`@- PIN 2.
`-I PI-II
`. = BLAST CELLS
`O = IATURE IVELOID CELLS
`
`.nO
`
`rm--
`
`antibody bound to leukemia cells. However, addition of
`saturating amounts of PM-SI in vitro showed that only
`about 25% of available antigenic sites had been occu-
`pied in vivo (Fig. 2). Plasma obtained 1 hr after the
`infusion was completed revealed high levels of free
`PM-81 antibody, as determined by the ability of this
`patient’s plasma to bind to normal PMNs in vitro.
`These studies suggested that antigenic modulation did
`not occur during the period of infusion of the PM-8]
`MoAb, since the number of antigenic sites was
`unchanged on cells exposed to PM-8l in vivo compared
`to pretreatment cells. Similar results were obtained
`during the subsequent infusions, though the percent-
`age of cells demonstrating in vivo antibody binding
`during the third and fourth infusions was lower than
`that observed in the first and second infusions.
`
`
`
`.a N
`
`
`
`CELLCOUNTl/JlI10' O
`
`17
`
`1!
`
`20
`
`19
`DAYS
`
`Fig. 3.
`
`Circulating blast and mature myeloid cell counts in patient 2 during MoAb treatment.
`
`Celltrion, Inc, Exhibit 100
`
`4 of 9
`
`Celltrion, Inc., Exhibit 1009
`
`

`

`MONOCLONAL ANTIBODY THERAPY OF AML
`
`1207
`
`to 5,000/ul. The following day, treatment with PMN
`29 produced a fall in the mature myeloid count from
`7,400/ul to 1,600 / u] and a decrease in the blast count
`from 3,400/ul to 1,400/ul. On the third day, PM-8l
`produced a fall
`in the mature myeloid count from
`6,600/ul to 400/ul and a fall in the blast count from
`5,300/ul to 1,400/nl. A return to pretreatment levels
`occurred within 1 day. Cells from a bone marrow
`aspirate performed on day 3 revealed no measurable
`cell-bound MoAb. During this treatment course, toxic-
`ity was limited to complaints of back pain during
`infusion of the PMN 6 antibody, which diminished as
`the rate of infusion was decreased.
`
`On day 15, infusion of PMN 29 (60 mg over 8 hr)
`decreased the mature myeloid count from 12,000/ul to
`4,100/ul and the blast count from 4,300/ul to 2,200/
`[.tl. PM-8l was then administered over 8 hr (60 mg)
`and 24 hr (210 mg). The blast count remained in the
`range of 1—2.000/ul, while the mature myeloid count
`dropped to 0 and remained low for nearly 24 hr. No
`toxicity, other than mild fever, was observed during
`this treatment.
`In View of the prolonged effect on
`reduction of the blast cell count, a larger dose (600 mg)
`of PM-81 was given by continuous infusion (4 mg/hr)
`over 7 days, starting on day 62 (with interruptions for
`blood transfusions) (Fig. 3). A fresh bottle of MoAb
`was prepared daily for this longer infusion. The mature
`myeloid count fell from 17,300/ul to 180/14] and the
`blast count from 4,500/ul to 1,800/ul. On day 65,
`when the infusion had been stopped for blood transfu-
`sion, major increases in both mature myeloid and blast
`counts were noted, but these counts fell again upon
`resumption of the infusion. On day 68, after therapy
`
`l 1,400/ul
`the blast count peaked at
`was stopped.
`level. The
`before falling again to the pretreatment
`serum LDH concentration tripled (200—600 [U/liter,
`isoenzymes 2 and 3) during each of these treatments.
`During the infusion of PM—81, MoAb could not be
`detected in the patient’s plasma or bound to cells from
`the peripheral blood or bone marrow. On day 64. a
`bolus of 125l-labeled PM~81 (0.25 mg PM-81, 7 x 108
`cpm/mg) was injected and serial measurements of
`plasma and cell-bound counts were performed. Rapid
`decrease in free plasma radioactivity occurred (T'/2 =
`3.5 hr), with plasma counts still detectable at 24 hr but
`not at 48 hr. Peripheral blood cell-bound radioactivity
`was detectable only immediately after infusion of the
`“SI-labeled PM-8l.
`
`Patient 3
`
`This patient was treated with MoAbs PMN 29 and
`AML-2-23. Since this patient had no circulating leu-
`kemia cells at the time of therapy and only a small
`percentage of blasts in the bone marrow, we monitored
`normal blood cell counts and investigated the ability of
`the AML-2—23 antibody to bind to leukemia cells in
`the patient’s leukemic skin nodules. Reductions in the
`peripheral blood neutrophil count and monocyte count
`occurred during infusion of MoAb AML-2-23 (60 mg
`over 12 hr) (Fig. 4). Infusion of PMN 29 (70 mg over
`12 hr) produced a decline in the neutrophil count to
`500/;11. Several hours later, infusion ofAML-2-23 (60
`mg over 12 hr) had no effect on the blood cell counts.
`Unbound PMN 29 was detectable in plasma following
`the PMN 29 infusion, but free AML-2-23 was not
`present during or after the infusion. Serum LDH
`
`l:= PMN 29
`-= AML-2-23
`"-8 LDH
`
`
`
`
`
`CELLcomm”:x10-3
`
`(I/nI)H01
`
`Fig. 4. Circulating blood cell counts and serum LDH
`levels in patient 3 during MoAb treatment. The effect of
`treatment on PMNs (0) and monocytes (I) is plotted in
`relationship to each MoAb infusion.
`
`Celltrion, |nc., Exhibit 100
`
`5 of 9
`
`Celltrion, Inc., Exhibit 1009
`
`

`

`
`
`the quantity of AML-2-23 blocking factor
`that
`increased after the infusion (Fig. 6). Free plasma
`AML-2-23 MoAb was not detectable immediately
`after a 60-mg infusion. These results also suggested
`that soluble and cell-bound antigens (on skin leukemia
`cells) were capable of binding all of the infused MoAb.
`Plasma from patient 2 and from normal donors did not
`contain significant
`levels of blocking factors
`for
`MoAbs PMN 29 and PMN 6 (data not shown)—
`findings that are consistent with our ability to detect
`free MoAb PMN 6 and PMN 29 in the circulation of
`
`
`
`%CELLSPOSITIVE
`
`PM-flqtglml)
`
`concentrations (isoenzymes 2 and 3) increased from
`300 to 800 lU/liter after MoAb treatment (Fig. 4).
`Removal of one of the patient’s leukemic nodules
`during the last infusion and isolation of these cells in
`vitro revealed that detectable levels of leukemia cell-
`
`bound MoAb AML-2-23 were present. However, the
`binding sites were well below saturation (approxi—
`mately 20% of saturation). During infusions of both
`antibodies, the patient experienced myalgias, arthral-
`gias. and fever, none of which necessitated discontin-
`uation of therapy. During treatment with AML-2-23,
`microscopic hematuria was noted. However, no change
`in the serum creatinine or BUN occurred.
`
`Evidence of Plasma Inhibitors of MoAb Binding
`to Cells
`
`Plasma samples from the patients treated in this
`study were found to have varying amounts of PM-Sl
`inhibitors (Fig. 5). Binding studies of purified PM-81
`in the presence and absence of pretreatment plasma
`from patient 2 indicated that 2—10 ug/ml of PM—8l
`was neutralized by a plasma substance, while plasma
`from patients I and 3 and normal donors had no
`blocking activity. Plasma samples obtained during the
`treatment of patient 2 with PM-81 did not block
`PM-SI binding in vitro and did not contain detectable
`
`Effect of in vitro incubation of
`Fig. 6.
`plasma from patient 3 on binding of MoAb
`AML-2-23 to PMNs. Plasma samples
`obtained before lday 2. I) and during
`AML-2-23 infusion (day 4. 0. and day 5. A)
`and from a normal donor (Ell were incu-
`bated with varying concentrations of
`AML-2-23. The binding of these mixtures
`and purified AML-2-23 alone (0) to normal
`PMNs was then detected by the addition
`of FITC-GAM, and the percentage of posi-
`tive cells was determined by cytofluorog~
`raphy.
`
`
`
`%CELLSPOSITIVE
`
`BALL ET AL.
`
`Effect of in vitro incuba-
`Fig. 5.
`tion of plasma from patients 1—3 on
`binding of MoAb I‘M-81 to PMNs.
`Plasma samples obtained before
`MoAb treatment from patient 1 (0).
`patient 2 (A). and patient 3 (0) and
`from a normal donor (El) were incu—
`bated with varying concentrations of
`PM-81. The binding of these mix-
`tures and purified I’M-81 (A) was
`detected by the addition of FITC—
`GAM and the percentage of positive
`cells determined by cytofluorogra-
`phy.
`
`amounts of free PM-81 (as determined by binding to
`PMNs in vitro). These findings provided presumptive
`evidence that soluble factors and cell-bound antigen
`were capable of complexing all of the infused PM-81
`MoAb. In contrast, the plasma of patient 1, which did
`not contain inhibitors of PM-81, did have free MoAb
`following treatment with PM-8l.
`Analysis of plasma samples obtained from patient 3
`before and after an infusion of AML-2-23 revealed
`
`this patient after infusion of these MoAbs.
`
`Patient Antibody Production to Mouse lg
`
`Plasma from patient 2 was found to contain human
`antibody of the lgG class reactive with murine MoAb,
`first noted on day 62 of therapy. The titer (defined as
`
`1
`
`AML-2-23(uglml)
`
`Celltrion, Inc, Exhibit 100
`
`6 of 9
`
`Celltrion, Inc., Exhibit 1009
`
`

`

`MONOCLONAL ANTIBODY THERAPY OF AML
`
`1209
`
`
`
`the dilution that gave 50% maximal binding) of this
`antibody increased from 1:100 on day 62 to l:l,600 on
`days 73 and 77. This antibody reacted similarly with
`lgM MoAb PMN 29 and the control lgM MoAb, but
`did not react with lgG MoAb AML-2-23, suggesting
`that the patient’s antibody is directed mainly toward )1
`chain.
`
`DISCUSSION
`
`These studies represent the first reported clinical
`trials of (l) lgM class MoAbs, (2) MoAb therapy in
`patients with AML,
`(3) combinations of MoAbs
`directed toward different myeloid differentiation anti-
`gens, and (4) MoAbs directed to glycolipid antigens.
`No significant
`toxicity was observed in any of the
`patients. Perhaps more importantly, the reduction in
`circulating leukemia cell count and the serum rise of
`intracellular enzymes indicated that leukemia cell kill
`was achieved in these patients.
`MoAb therapy in mice has demonstrated that lgG
`(but not lgM) antibodies lead to an enhanced ability of
`phagocytic cells to mediate leukemia cell
`lysis,""17
`probably through antibody-dependent cellular cyto-
`toxicity (ADCC) or by opsonization of target cells.
`Although it
`is possible that,
`in the present study,
`MoAb coating led to transient sequestration of leu-
`kemic cells in extravascular sites, the depression of
`peripheral blood neutrophil and blast cell counts con-
`comitant with a rise in leukocyte LDH isoenzymes
`indicate that myeloid cell lysis did occur in vivo. The
`mechanisms of this cytolysis are at present unclear.
`They probably do not
`include direct complement-
`mediated lysis, since the MoAbs employed in this
`study did not mediate significant cell
`lysis in vitro
`using human complement. Furthermore, there is no
`evidence for the existence of lgM-specific Fc receptors
`on human neutrophils or monocytes.I8 On the other
`hand,
`lgM MoAb-mediated cytolysis may occur
`through fixation of C3b to complement receptors on
`phagocytic cells in the absence of activation of the
`subsequent complement components. Alternatively, or
`in addition, since these MoAbs bind to normal phago-
`cytic cells as well as leukemia cells, bridging of target
`and effector cells might have occurred without
`the
`participation of complement or Fc receptors.
`With the introduction of the concept that tumors
`may issue from a pool of proliferating and self-
`renewing cells representing only a fraction of the
`tumor,'9 and the increasing evidence that various fac-
`tors may infiuence the rate of proliferation and degree
`of differentiation in cells that have emerged from this
`pool, our findings gain greater significance. Thus,
`although it would be desirable to eliminate or greatly
`reduce the size of the tumor stem cell pool (as presum-
`ably occurs in that
`fraction of patients achieving
`
`long-term remissions or cures with chemotherapy), it is
`possible that treatments eliminating cells at stages
`beyond the stem cell
`level could reduce the disease
`morbidity or could induce kinetic events in the stem
`cell pool, thereby rendering it more sensitive to chemo-
`therapy.
`in vivo binding to
`Following infusion of MoAbs,
`leukemia cells, although subsaturating, was demon-
`strated in each patient. Moreover,
`in patient 1, the
`inability to saturate leukemia cells occurred despite
`the presence of 0.5—1 ug/ml of free PM-8l, suggesting
`the existence of different subpopulations of leukemia
`or normal myeloid cells with diverse affinities for the
`antibody. Differences were demonstrated between
`patients in the quantity of free plasma MoAbs subse-
`quent to administration. Free plasma MoAbs PMN 6
`and PMN 29 were demonstrable in patient 2, but
`PM—8l was not detectable during or after infusion in
`this patient. Free PMN 29, but not AML-2-23, was
`detectable in patient 3. Clearance of free MoAb could
`occur by binding to cells, binding to soluble factors,
`catabolism, or by immunologic elimination resulting
`from the development of human anti-mouse antibody.
`Binding to cells and to soluble factors is the most likely
`reason for the rapid clearance of MoAbs observed in
`these studies. Elimination due to anti-mouse antibody
`is likely to have occurred only during the later infusion
`given to patient 2