`
`LindenDrive
`
`
`Edited by R. W. Baldwin and E. IVIihich
`
`
`_T
`
`Volume39 Number6 1994
`
`
`Original articles
`
`alignant mesothelioma,
`TIL, CD3, TGFB,
`Interleukin-6
`
`Bielefeldt-Ohmann H, Fitzpatrick DR, Marzo AL, Jarnicki AG, Himbeck RP, Davis MR,
`Manning LS, Robinson BWS: Patho- and immunobiology of malignant mesothelioma:
`
`characterisation of tumour infiltrating Ieucocytes and cytokine production in a murine model 347
`
`
`Tumor—induced
`Sahasrabudhe DM, Dusel JC: Effect of murine interferon edit on tumor-induced suppressor
`function
`360
`immunosuppression,
`IFNou'B
`
`
`
`
`
`
`ieloid leukemia, CD33,
`Immunotoxin, Gelonin,
`M195, Huh/I195
`
`
`
`Radioiodinated
`monoclonal antibodies,
`F(ab')2 fragments,
`Chimeric antibodies,
`
`Biodistribution
`
`
`McGraw KJ, Rosenblum MG, Cheung L, Scheinberg DA: Characterization of murine and
`humanized anti-CD33, gelonin immunotoxins reactive against myeloid Ieukemias
`357
`
`Westera G, Rtilicke T, Smith A, Duewell S: Differences in biodistribution of the anti-
`(carcinoembryonic antigen) murine monoclonal antibody CE-25, its Flab’)2 fragment and its intact
`mainly human chimeric form CE 4-8—13. Dependence on tumour size and amount of antibody
`injected 375
`
`
`
`
`
`
`
`Squamous cell
`Carcinoma, TIL, Tcells,
`NK cells, MHC class I,
`p53
`
`Hald J, Rasmussen N, Claesson MH: In vivo infiltration of mononuclear cells in squamous cell
`carcinoma of the head and neck correlates with the ability to expand tumour-infiltrating Tcells in vitro
`and with the expression of MHC class I antigens on tumour cells
`383
`
`
`Immunotherapy,
`Haagen l-A, Lau WBM do, Bast BJEG, Geerars AJG, Clark MR, Gast BC de: Unprimed CD4+ and
`Bispeciftc monoclonal
`CD8+ T cells can be rapidly activated by a CD3><CD19 bispecific antibody to proliferate and become
`antibodies, T cell
`cytotoxic
`391
`activation, Cytotoxicity
`
`
`
`
`
`Human IgM, Clinical
`Rosenblum MG, Levin B, Roh M, Hohn D, McCabe RéThompson L, Cheung L, Murray JL:
`pharmacology,
`.
`Clinical pharmacology and tissue disposition studies of 1 1I-Iabeled anticolorectai carcinoma human
`lelabeied monoclonal
`
`monoclonal antibody LiCO 16.88 397
`antibodies
` 1
`
`Contents continued on the back cover Springer international
`
`262 Cancer Immunol Immunotheri ISSN 0340—7004 CII MDN 39 (E3) 347-420 (1994)
`Printer-i nrt acid-frat: nannr
`____—
`IMMiLJ'NCG-EN 2142, pg. 1
`Phigenix v. Immunogen
`IPR2014-00676
`
`m 4
`
`A
`
`IMMUNOGEN 2142, pg. 1
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`h
`
`Cancer Immunol Intnnlnother(1994) 39: 367—374
`
`
`
`
`
` Cancer
`I
`
`mmunology
`mmunotherapy
`© Springer—Voting 1994
`
`This material may be protected by Copyright law (Title 17 U.S. Code)
`
`Characterization of murine and humanized anti-CD33,
`gelonin immunotoxins reactive against myeloid leukemias
`
`Kimberly J. McG’i‘awl, Michael G. Rosenbluml, Lawrence Cheung3, David A. Scheinhergl
`l Memorial Sloan Kettering Cancer Center, New York. NY 1002l. USA
`3 Immunopharmacology Section. Box 41. Department of Clinical lmnutnology and Biological Therapy. M. D. Anderson Cancer Center.
`1515 I-Iolcombc Blvd. Houston. TX 77' 030. USA
`
`Received: 9 March l994/Accepted 26 August 1994
`
`Abstract. M195 antibodies recognize CD33, an antigen
`present on acute myeloid leukemia blasts as well as some
`myeloid progenitor cells, but not on the ultimate hemato—
`poietic progenitor stem cell.
`lmtnunotoxins (IT) reactive
`with human myeloid leukemias were constructed by con—
`jugating gelonin,
`a
`single—chain ribosome—inactivating
`protein, to marine and genetically engineered, humanized
`M195
`antibodies via
`an N—succinimidyl'3—(2—pyridyl—
`dithio)—propionate linkage. No losses of gelonin cytotoxic
`activity or M195 binding activity were observed after
`conjugation of up to two toxin molecules per antibody.
`Toxin conjugates displayed specific, potent
`toxicity for
`CD33+ cells. The murine and humanized IT were not toxic
`to CD3} cells and were 600 and 4500 times tnore potent,
`respectively, than free gelonin in inhibiting CD33+ HL60
`cells. Treatment of HL60 cells with 1 uglml l-luMl95—ge-
`lonin resulted in more than 1000 times lower colony for—
`mation; normal bone marrow mononuclear cell colony-
`forming units treated with HL1M1957IT were reduced by a
`factor of 10. HL60 leukemia ceils could be effectively
`purged from an excess of normal bone marrow cells. Exe
`posure of target cells to IT for as little as 30 min was as
`effective as continuous exposure of IT'for up to 6 days.
`However, measures of the efficacy of the iminunotoxin
`were directly related to the length of time of observation
`after IT exposure and were inversely related to cell con
`centration. M195-gelonin immunoconjugates are potential
`candidates for therapeutic Lise in in vivo or ex vivo bone
`marrow purging of myeloid leukemias.
`
`Key words: Myeloid leukemia 7 CD33 7 Immunotoxin ~
`Gelonin 7 M195 — HuM195
`
`These studies were supported in part by the Lucille P. Markey
`Charitable Trust. ACS Grant No. IMSSI. NIH POlCA33049. NIH
`ROICA55349. Research conducted. in part. by the Clayton Foundation
`for Research. David A Sciieinberg is a Lucille R Markey Scholar
`
`Correspondence to: D. A. Scheinberg
`
`Introduction
`
`Immunoioxins (1T) are a class of proteins that consist of a
`monoclonal antibody (mAb) covalently linked or geneti-
`cally fused to a toxic molecule and are thus able to direct
`potent cytotoxicity to particular cells [41. 22]. Obstacles to
`effective therapeutic use of itnmunotoxins for cancer in~
`clude (a) lack of suitable tumor~specific targets that are not
`also found on other vital notrtumor cells [39]; (b) loss of
`tOxin potency or mAb activity after conjugation [15];
`(c) unwanted cytotoxicity to nontarget cells and tissues
`resulting from nonspecific internalization of the IT [22];
`(d) immunogenicity of the IT [10. 21. 27]; (e) pharmaco-
`logical inability to target tumor sites adequately [10, 20].
`CD33 provides a useful
`target antigen for therapy of
`myelogenous leukemias. as it is expressed on the cell sur—
`face of more than 80% of leukemia isolates from patients
`with myeloid leukemia with an average antigen density of
`10000 sites/cell
`[1. 18, 32, 38]. In addition, rapid inter-
`nalization occurs upon binding of mAb to CD33 both in
`vitro and in vivo [14, 33]. CD33 is also found on normal
`granulocyte/monocyte-colony-forming units
`(CFU—GM),
`some burst—forming colonies and a fraction of the more
`primitive progenitors [1. 18, 32, 38]. However. CD33 is not
`found on tissues outside of the hematopoietic system nor on
`the normal pluripotent hematopoietic progenitor stem cell.
`My9, a murine IgGZB mAb reactive with CD33. has been
`conjugated to a chemically modified ricin toxin and de—
`monstrates potent specific cytotoxicity in vitro [23, 31].
`Both M195, a murine monoclonal
`IgG2a antibody, and
`HuM195. a humanized version of M195 constructed by
`genetically grafting the marine CDR regions to a human
`IgG1 framework and constant regions. are reactive with
`CD33 [8. 12. 33, 38]. The humanized antibodies may be
`advantageous because of their reduced immunogenicity.
`higher avidity, and longer serum half lives [7. 8, 12].
`Both murine and humanized M195 are now in clinical
`trials for the treatment of myelogenous leukemias.
`”ll—
`radiolabeled M195 is capable of killing up to 1013 leukemia
`cells in patients with refractory or relapsed leukemias [35].
`However, because of the long—range cytotoxicity of the
`
`‘——_————_—_—#
`
`IMMUNOGEN 2142, pg. 2
`Phigenix v. lmmunogen
`lPR2014-00676
`
`IMMUNOGEN 2142, pg. 2
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`$4.]1.11":L21_'_!H“1"
`
`368
`
`conjugated nuclide (approximately 50 cell diameters),
`killing of normal bystander cells occurs as well, requiring
`bone marrow transplantation at high doses. Use of an anti—
`CD33 IT may avoid this problem by killing only those cells
`bearing the CD33 antigen.
`Gelonin is a 30-kDa single chain ribosomeinaetivating
`protein isolated from the seeds of Gelonium nrnltiflortnn,
`which irreversibly inactivates the ribosomal 60$ subunit by
`cleaving the N—glycoside bond of adenine in a specific se—
`quence of rRNA and thereby prevents protein elongation
`[2, 37]. Gelonin has several advantages over other ribo-
`someuinactivating proteins currently being investigated.
`Unlike the dual-chain type, such as ricin [4] and abrin [36],
`which contain a B chain reactive with cell—surface carboe
`
`hydrates, gelonin is a biochemically stable single chain [22,
`29, 36]. It lacks a galactose—specific lectin domain, which is
`responsible for non—specific binding and toxicity to cells.
`As a result,
`free gelonin is much less toxic to intact
`mammalian cells in vitro and in vivo than the heterodirneric
`toxins. Despite this relative safety, in a ceil~free rabbit ref
`ticulocyte translation assay, gelonin demonstrates nearly
`equal activity to heterodimeric toxins [6, 26]. Attempts
`have been made to prevent the nonspecific binding of the
`B chain of heterodimeric toxins by separating the A and
`B chains or by blocking of galactose-binding sites [2], 3 l ].
`These methods are technically difficult, variably effective,
`or sometimes incomplete; hence the IT may contain trace
`contamination with active B chain that can result in a less
`
`selective immunoconjugate and potential toxicity in vivo.
`Residual membrane—binding activity of the ricin B chain
`may be necessary for translocation of the A chain [17]. In
`addition,
`the larger heterodimers may also be more irn—
`munogenic than the single—chain toxins.
`Thus, gelonin presents a potential advantage over the
`dual-chain toxins and was selected for this study. In this
`paper, we describe immunotoxins constructed by chemical
`conjugation of mAb reactive against CD33 antigen with the
`plant toxin gelonin.
`
`Materials and methods
`
`Cell cnlmre. Cell lines HL60 (acute myeloid leukemia, CD330. U937
`(monocytic leukemia, CD33+), Raji and Daudi
`(B lineage Burkitt‘s
`lymphomas. CD331 Molt4 (T lineage lymphoma. CD33-) and
`SKLY16 (B lineage lymphoma, CD33), were maintained in culture
`using RPMIvlfitiO medium supplemented with 10% Serum Plus (.IRH
`Bioseiences), 5% heatiinactivated newborn bovine serum (Armour
`Pharmaceuticals). non-essential amino acids. penicillin and strepto—
`myciti.
`Transformed murinc fibroblast cell litres NIH—3T3 and AL67 [24],
`the latter expressing the transfected CD33 gene, were maintained irr
`culture similarly.
`
`Antibodies. mab Ml95 and HLrM195 were prepared as described [8, 32,
`38]. Highly purified HuM195 and HuGl Fd79 [ll], a genetically en?
`ginecred human IgGl construct reactive with a herpes simplex virus
`antigen not found on HL60 cells, were the generous gifts of Man Sung
`Co, Protein Design Labs. Mountain View, Calif.
`
`SPDP conjugation. A threefold molar excess of Nisuccinimidyl 37(27
`pyridyldithio)Apropionate
`(SPDP), prepared in dry dimethyl-for-
`mamide, was added to l0 mg Ml95 or HuMlQS in phospliateibuffered
`
`saline pH 7.4 (PBS) and incubated for 30 min at room temperature.
`Excess SPDP was removed on a Sephadex (3-25 column cquilibruted
`with 0.1 M phosphate buffer pH 7.0 containing 0.5 mM EDTA.
`
`Modification quelonin with 2-irninaflriolane. Z—Iminothiolane in 0.5 M
`triethanolamine hydrochloride (TEA/HCI). pH 8.0. was added to 10 mg
`gelonin in 60 mM TEA/HC].
`1 mM EDTA. pH 8.0, and incubated for
`90 nrin at 4 EC under nitrogen. Excess 27irninothiolane was removed
`by gel filtration on a Sephadex G-25 column equilibratcd with 5 mM
`BlSTRISlacetale buffer pH 5.3 containing 50 mM NaCl and 1 mM
`EDTA.
`
`Conjugation ofMl795 with gelonin
`
`SPDPimodified M I95 in 100 mM sodium phosphate buffer. 0.5 mM
`EDTA. pH 7.0, was mixed with a fivermolar excess of 2—iminothiow
`lane-modified gelonin. The pH was adjusted to 7.0 with 0.5 M TEA!
`HCl. pH 8.0. and the mixture was incubated for 20 h at 4 DC under
`nitrogen. To stop the reaction,
`iodoacetamide was added to a final
`concentration of 2 mM and incubated for I h at room temperature.
`
`Purification. The reaction tnixtttre was filtered on a Sephacryl 5—300
`gel filtration column equilibrated with 20 mM TRIS. 50 mM NaCl,
`pH 7.4. to separate the antibody and antibody—gelonin conjugates from
`the free gelonin. The fractions containing immunotoxin and unreacted
`antibody were pooled and then loaded on a Cihacron-blucASepharose
`CL—(iB column equilibrated with TRIS buffer
`to remove the un—
`conjugated antibody. Purified innnunotoxin (MIQSAIT) was eluted with
`20 mM TRIS buffer containing 2 M NaCl. pH 7.4. Collected fractions
`were dialyzed against PBS.
`
`Flair-Luminary assays. Cells were washed and resuspended in RPML
`1640 medium, with 10% Serum Plus, non—essential amino acids, pee
`nicillin, streptomycin. and 2% rabbit serutn (Pei Freeze). Samples
`containing 500000 cells in a final volume of (HS ml were incubated
`for l h on ice in the presence of primary antibody. Cells were washed
`twice,
`incubated for
`l h on ice with secondary fluoresceiniisoi
`thioeyantue(FlTCHabelled antibody [goat anti-(mouse lg) or goat
`anti—(human Igjj, washed twice, and fixed with 0.5% paraformalder
`hyde. The FITC fluorescence intensity was measured on an EPICS
`Profile 1] flow cytometer [32].
`
`Enzymeefinked innnnnosorbent marry {ELISA}. Primary antibody was
`added to 96—well plates Containing adherent AL67 or NIH-3T3 cells
`and allowed to bind for l
`h at room temperature. CD33 is over-
`expressed and is not modulated quickly off of the surface of AL67
`cells. Excess antibody was removed by washing the cells three times.
`Alkalinc-phosphatase—labelled secondary antibody was added and al—
`lowed to incubate for l h at room temperature. Celis were washed
`three times and 100 ul substrate solution p—nitrophenyl disodium
`phosphate was added. Absorbance at 405 nm was measured using a
`Fisher Biotek microplate reader after a 10-
`to 20-min incubation at
`37 DC.
`
`Rabbit reticulocyte lysnre translation assay. Tlie,.functional activity of
`gelonin and gelonin-containing irnmunotoxin ”was assayed by a eelli
`free translation inhibition assay kit (Gibco—BRL) as described by the
`manufacturer:
`
`Inhibition of [3er‘lrrnridnm [iHjiencina and -“H-lrtbeiied amino acid
`incorporation. Samples containing 100 [.11 cells were washed and in,
`cubated at 37 “C in 967well plates in the presence of 50 ul antibody.
`conjugate, or toxin. After an incubation time of 377 days, 50 ul
`[0 uCi/ml tritiated thyrnidine, leucine. or amino acids was added to
`each well and allowed to incorporate for 5—6 h. A 507ttl aliquot of
`triehloroacetic acid was added to precipitate protein for [3H]7leucine
`and JIrl-labelled amino acid incorporation experiments. Cells were
`harvested with a Skatron semiautomatic harvester and assayed in a
`Packard scintillation counter.
`
`IMMUNOGEN 2142, pg. 3
`Phigenix v. Immunogen
`|PR2014-00676
`
`IMMUNOGEN 2142, pg. 3
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`3~[4,5edtmelftj-dthit’i‘zolijvfel5115p]?cnyft‘cfrfl:0.litun bromide (ttziazoyl
`blue: MTT) assays. Samples containing 100 ttl cells were washed and
`incubated at 37 DC in 96-well plates in the presence of 50 [41 antibody,
`conjugate, or toxin. After an incubation time 01°37? days. the plate
`was centrifuged for 5 min at 1000 rpm. MTT, diluted in PBS, was
`added to each well and incubated for 4 h. Plates were washed. The
`formazan prodtict was solubilized with 0.04 M HCl in propanfliol and
`quantified spectrophotometrically at 570 nm.
`
`Competition assays. HL60 cells at 1x105 cells/ml were incubated for
`1 h on ice in the presence of excess HttM195 or Htth79. HuM195—1T.
`at a concentration capable of killing approximately 70% of cells, was
`then added. Cells were incubated for 90 h at 37 “C then assayed by
`trypan blue exclusionio‘r [3H]thymidine incorporation.
`Time course studies. HL60 cells at 0.67x 105 cellsfml were incubated
`with l-luM195—1T. At various times cells wee washed twice and ref
`suspended in an ITifree tnedittm. On day 6 the cells were plated onto a
`96—well plate and analyzed by trypan blue exclusion or [3Hjthymidine
`incorporation.
`
`Clrmogem'c growth. HL60 cells were treated with HuM19571T and
`allowed to incubate for 24 h at 37 °C. Cells were then washed and
`plated in 1 ml 0.3% agarose, RPM1-1640 medium. 8.33% newborn
`bovine serum. 18.33% fetal calf serum (FCS). non-essential amino
`acids. penicillin and streptomycin. Plates were incubated for 14 days at
`37 °C.
`
`Fourteen-day CFU-GM. Fourteen—day CPU—GM assays were per—
`formed essentially as described [16] using= various cell concentrations
`plated onto l—inl agarose dishes supplemented with granulocyte—col-
`coy-stimulating factor.
`grann1ocyte/macrophage-colony—stimulating
`factor and interleukinfi. Growth factors were the generous gift of
`Janice Gabrilove, MSKCC. Assays were clone in duplicate or quad-
`ruplicate and each experiment was repeated three times. The plating
`efficiency was approximately 0.1%.
`
`Pinging of HLoU from excess nornm! bone marrow. Bone marrow
`aspirates were obtained from normal donors according to Memorial
`Sloan Kettering Cancer Center IRB protocols. Mononttcicar cells were
`collected by Ficoieraqtie sedimentation, washed. and gamma—ina—
`diated with 8 Gy. Marrow cells were divided into aliquots in 96-well
`plates at a final concentration of 1x10“ cells/ml, HL60 cells at a final
`concentration of 0.667x105 cells/ml and HuMl95—1T at various con
`centrations were added to the plates. After a 6—day incubation at 37' DC.
`cells were assayed for [3Hjtliyltiidilte incorporation.
`
`Results
`
`Conjugation and purification
`
`Both M195 and HttM195 were conjugated with gelonin and
`purified as described in the Materials and methods. As
`shown in Fig. 1, purified M195 antibody migrates on the
`sodium dodecyl sulfate/polyacrylarnide gel electrophoresis
`(SDS-PAGE) as three major protein bands representing the
`glycosylated and carboxyterminalrmodified forms of M195
`[1 3]. The final pttrified conjugate was also fottnd to contain
`three major protein bands as shown in lane C. Since,
`electrophoretic analysis alone could not confirm whether
`the final purified immunotoxin contained any unconjugated
`antibody, analysis of the final
`immunotoxin preparation
`was done by Western blot analysis using anti-gelonin rabbit
`polyclonal antisera to confirm the presence of gelonin in
`each of
`the major Coomassie-Stained bands (data not
`shown).
`
`369
`
`i.
`
`- 14—195
`z r.
`{.3 - RXN Hixture
`fl???
`- Purified
`11—195 Gelonin
`- Gelonin
`
`.
`
`.
`
`Fig. 1. Sodium dodecyl sulfate gel electrophoretic fractionation of
`Ml95 and the immunotoxin Ml95-1T under non-reducing conditions
`with a gradient of 5%e20b7i7 acrylamide. Lanes from left
`to right:
`purified mouse M195: reaction mixture containing M195 and gelonin;
`purified M195-gelonin immunotoxin; purified gelonin
`
`Binding spect'ficint and titer of IT
`
`The binding specificities of M195-1T and HuM195-IT were
`tested against a variety of CD33+ (U937, HL60, AL67) and
`CD33’ (SKLY16. Raji, Molt4, Daudi) cell lines using flow
`cytometry or ELISA. The conjugates retained the same
`specificity for CD33-expressing cells as the unconjugated
`antibodies.
`
`Different batches of conjugates containing approxi-
`mately one or two gelonin molecules per mAb showed no
`loss of binding titer as compared to unconjttgated HuM195
`(Fig. 2A, B) or M195 (data not shown). However. a batch
`of HuM195 conjugated with an average of three gelonin
`molecttles per mAb, as determined by SDS—PAGE. des
`monstrated a lower binding titer against both HL60 cells, as
`assayed by indirect flow cytometry, and AL67 cells, as
`assayed by ELISA (data not shown). Therefore, all addi-
`tional experiments presented in this paper used batches of
`gelonin conjugated to antibody at a ratio of approximately
`1
`:
`1 unless otherwise stated.
`
`Biological activity of toxin.
`
`The ability of HttM195—IT and gelonin to inhibit translation
`in a cell-free system was assayed using a rabbit reticulocyte
`lysate translation assay. Both the immunotoxin and free
`gelonin demonstrated a similar 50% inhibitory concentra—
`tion (IDsu) in the range of 12716 pM, demonstrating that
`conjugation of gelonin to the antibody did not alter the
`activity of the toxin (Fig. 3).
`
`Cytotoxiciry 0f immtmoconjugtttes. Both M195-IT and
`HuM195—IT were tested for their ability to kill CD33+ and
`CD3} cells in comparison to free gelonin. Cytotoxicity
`was detennined by a variety of methods including inhibi-
`tion of incorporation of tritiated amino acids into tri—
`chloroaceticacid—precipitable protein.
`inhibition of DNA
`synthesis, trypan blue excl usion, MTT, and clonogenic as—
`says (not all shown). M195—IT had an 1D50 of approxi—
`mately 400 pM. which was approximately 600 times more
`potent than free gelonin (Fig. 4). HuM19571T, which has a
`higher affinity for CD33 than M195, had an [D50 of 15 pM,
`which was 4500 times more potent than the 1D50 of free
`gelonin (Fig. 5). Different lots of HuM195—IT displayed
`
`
`
`IMMUNOGEN 2142, pg. 4
`Phigenix v. Immunogen
`|PR2014-00676
`
`IMMUNOGEN 2142, pg. 4
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`0 Hu-MiBE-Gelunin
`E! Gelunin
`
`100
`
`o
`
`x"
`
`Inhibition 8
`Percent
`
`10—1510'1410—1310‘1210‘1110‘10 10‘9 104' 10*7 10—5
`Protein Concentration. Molar
`
`Fig. 3. Rabbit retictrlocyte lysate translation assay. Increasing amounts
`of free gclonin toxin (El) 0r HuM 195-1T (0) were assayed for activity
`in a ceil—frec lysate translation inhibition assay. HLiM1957IT final
`concentrations
`ranged from 5 pglnil
`to 5
`rig/ml. Gelonin final
`concentration ranged from 1 pg/rnl
`to |
`lug/ml
`
`100
`
`J:-a)00DO0
`
`E
`
`
`
`PercentageAminoAcids
`
`Incorporated
`
`1E-11
`
`1E-1O 1E-GQ 1E-08
`
`1 ED?
`
`1E-06
`
`1E-05
`
`Toxin Concentration (M)
`
`Fig. 4. Inhibition of protein synthesis in live cells by gelonin and
`M195-IT on HL60 cells. HLEO cells at a final concentration of l X It)"
`cells/ml were incubated for 3 days at 37 °C in the presence of M19571T
`(I) and gclonin (El). Levels of protein synthesis were determined by
`5 h incorporation of tritiated amino acids into trichoroacetic-acid-
`precipitable protein. Ml95-IT final concentrations ranged from 5 rig/
`ml
`to 4 tlg/rnl. Gelonin final concentration ranged from 0.5 ttg/rnl
`to
`50 pg/rnl. The data are representative of fotrr experiments
`1'
`
`Because of the increased potency shown above and the
`theoretical advantage of reduced immunogenicity of the
`humanized form of the M195 in vivo, all further experi-
`ments were conducted using the HuMI95—IT.
`HuMl95-IT inhibited the clonogenic growth of HLGO in
`a dose-dependent manner. Incubation of cells with 10 pg/
`ml and l
`ttg/ml I-luMl95~lT for 72 h decreased colony
`formation from 8225 colonies/105 cells plate-d to 5 and 13
`colonies/105 cells plated respectively (Fig. 6). Incubation
`with 0,1 rig/ml HuMlQS—IT and trnconjugated antibody did
`not significantly inhibit growth.
`
`IMMUNOGEN 2142, pg. 5
`Phigenix v. Immunogen
`|PR2014-00676
`
`A
`
`4
`
`3
`
`2
`
`1
`
`.01
`
`.1
`
`1
`
`10
`
`0.
`
`001
`
`.01
`
`.1
`
`1
`
`1o
`
`B
`
`5
`
`4
`
`:3
`
`2
`
`1 o.
`
`001
`
`370
`
`E’r:
`M
`.C
`
`0 f
`
`i
`G.)
`o.
`
`Eo
`E
`
`_
`d.)
`
`CC(
`
`Uc
`
`(.1
`
`x 3a
`
`EM g
`
`w1::A:
`’‘s‘cJL—
`
`(pg/ml)
`Antibody Concentration
`Fig. 2A, B. Binding titer and specificity of HuM l957IT on cell lines.
`HL60. U937. or Molt-i cells at a final concentration of LSX 106 cells!
`ml were incubated on ice for | h with either HuGl M195 or HuGl
`M195—IT a a final concentration range of 0.08 pg/rnl
`to 10 ttg/rril,
`Mean peak fluorescence intensity (y axis) versus rnAb or immunotoxin
`(1T) concentration (.r axis) was measured using an EPICS Profile II
`cytometer. A HLoG binding by HuMl937IT t.) or by HuMlQS (O),
`B U937 binding by HLIMIQSJT t.) or by Huh/1195 (O): Molt4
`binding by HuM |957IT (I) or by I-luM195 (A)
`
`different levels of cytotoxicity with some lots having an
`ID70 Of 5 pM. There was a slow loss of potency over time
`(months) suggesting either reduction of the intermolecular
`linkage or denaturation of the IT. Because of the vari—
`abilities, accurate comparisons of potency between the
`marine and humanized conjugates are not possible.
`Cytotoxicity, as determined by 3Hrlabelled amino acid
`or [3H]thymidine incorporation, was confirmed to result in
`cell death by examining parallel wells by trypan blue
`analysis or similarly treated cells as measured by MTT (not
`shown).
`Non-specific cytotoxicity was not observed with the IT.
`Even at the highest concentration of irnrnunotoxin used,
`typically 10~ 16 rig/ml, no cytotoxicity was observed when
`either conjugate was incubated with the CD33-negative cell
`lines Raji. Moltél, Daudi, and SKLYlo (not shown).
`
`IMMUNOGEN 2142, pg. 5
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`100
`
`75
`
`50
`
`through the CD33—antigenebinding site and not through the
`Fc region or other non—specific binding sites on target cells.
`
`371
`
`Gelonin
`
`Efibcts of target cell concentration
`
`HuM195 IT
`
` 2 5
`
`1E-13 1E-12 1E-11
`
`1E-10 1E109 1E-03 1E-07 1E-06 1E-05
`
`
`
`
`
`PercentageThymidineIncorporated
`
`In order to determine the effect that cell concentration may
`have upon the efficacy of the HUM] 95—1T. HLoO cells were
`serially dilttted and incubated in the presence of a single
`concentration of IT at 2 pg/ml. The immunotoxin was most
`effective at low cell concentrations (Fig. 7'). At higher cell
`concentrations, the IT lost potency. This was not due to an
`excess of cell surface binding sites over IT molecules as.
`even at
`the highest
`cell
`concentrations,
`there were
`100—1000 more molecules of IT than available binding
`sites. In other experiments. serial dilutions of IT and ge-
`lonin were incubated with three different concentrations of
`
`HL60 cells. Both the IT and gelonin were more potent. at
`lower cell concentrations (data not shown), Experiments
`looking for transferable cell—free inhibitory factors secreted
`or released by target cells in high concentrations were
`negative (not shown).
`
`Killing of leukemia tc'n'gets cotitmniimting noriitcn' bone
`HHU'I'OH'
`
`The dependence of killing on cell density raised the issue of
`the efficacy of the IT in the presence of large numbers of
`non-target cells. Therefore.
`to determine whether the IT
`was able to kill HL60 cells in the presence of excess
`numbers of CD33— cells. HL60 cells were mixed with a
`t5fitirnes excess of irradiated normal bone marrow cells.
`This ratio simulates that which might be found in a typical
`marrow in early relapse contaminated with low levels of
`leukemia cells. The presence of the bone marrow had
`minimal effect upon the cytotoxicity of the IT (Fig. 8).
`
`Efifet'ts of length Qfetpostrre oflT with target cells. CD33
`modulation and internalization begin within minutes of
`mAb binding in vitro and in vivo [14, 33. 38]. HL60 cells
`were incubated in the presence of HuMIQS—gelonin at 1 pg/
`ml for 30 min, 1 h, 4 h, and 6 days and then washed twice to
`
`Toxin Concentration (M)
`
`Fig. 5. Inhibition of DNA synthesis in live cells by gclonin and HuGI
`M195-1T on HL60 lines. HLoO cells at a final concentration of 3X 104
`cellslnil were incubated for 5 days at 37 °C in the presence of H116]
`M19571T t.) or gelonin (0). Levels of DNA synthesis were
`determined by 5 h incorporation of tritiatcd thymidine. HuGl M195-
`IT final concentrations ranged from 0.2 Jig/ml to 4 tlglml. Gelonin final
`concentration ranged from 0.5 ttglml
`to 50 uglnil. The data are
`representative of five experiments
`
`l ug/ml and 10 pg/ml inf
`I-IuMl95—IT at 0.25 lag/ml,
`cubated with bone marrow mononuclear cells resulted in
`60%—90% decreases in colonyeforming units. Gelonin at
`the same molar concentration as 1 pg/ml IT resulted in an
`approximately 70% decrease in colonyrforming units.
`
`Specificin'for CD33 sites
`
`To confirm further the specificity of cytotoxicity of the
`HuM195—1T for CD33+ cells and its requirement for bind-
`ing to the CD33 antigen itself, HL60 cells were incubated
`on ice in the presence of an excess of unconjugated
`l-luM195 or HuGle79. an isotype control antibody. Pref
`incubation of HL60 cells with 50 uglml HuMI95 was able
`to inhibit more than 85% of the HuM195—IT cytotoxic ac-
`tivity, whereas HuGle79 did not block killing. This
`confirmed that specific targeting of leukemic cells was
`
`(M)
`
`HuG1M195-1T
`
`l
`
`0.1
`
`1
`
`10
`
`100
`
`1000
`
`10000
`
`100000
`
`HL60 Colonies per 100.000 Cells nlated
`
`Fig. 6. Effect of l—IuGl M195 upon HL60 colony
`formation. HLoO cells at 6.67 X 104 cells/ml were
`
`incubated in the presence of various concentra-
`tions of HuGI M195 for 3 days. Cells were then
`washed and plated in 0.36% agarose, 20% fetal
`calf serum. 10% Serum Plus. Colonies were
`assayed after a i4rday incubation. The data are
`representative of two experiments
`
`
`
`IMMUNOGEN 2142, pg. 6
`Phigenix v. Immunogen
`|PR2014-00676
`
`IMMUNOGEN 2142, pg. 6
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`various concentrations of HuMlQSeIT were harvested after
`3, 4, and 5 days of exposure. The median lethal dose (LDso)
`at 5 days (0.3 nM) was 30 times lower than the LD50 at
`4 days (8.9 nM) and 40 times lower than the LDsn at 3 days
`(12 nM). Therefore, although the initial event necessary for
`cytotoxicity occurred within 30 min, the subsequent events
`leading to cell death required considerable time to develop
`fully.
`
`Discussion
`
`Major obstacles to the effective therapeutic Lise of IT in
`humans include immunogenicity of the IT, toxicity to cells
`that non~specifically internalize the IT. and difficulty in
`delivering sufficient IT to tumor sites [10, 15, 19, 20—22.
`39. 41]. The HuMlQS—gelonin iinmunotoxin may bypass
`seine of these difficulties by use of a CDR—grafted non-
`itnmunogenic mAb that has already demonstrated rapid.
`efficient, specific, and saturable targeting to leukemia cells
`in humans in vivo. and by conjugation of that mAb to a
`smaller, less toxic RIP.
`
`M1957gelonin conjugates showed specific and potent
`cytotoxicity for CD33+ leukemia cells, required short ex
`posure times for activity, and were unaffected by the
`presence of excess
`irradiated bone marrow. The
`lin-
`munotoxin conjugates containing fewer than three gelonin
`molecules per antibody maintained complete biological
`activity of the mAb, as measured in fiow-cytometric assays,
`and toxin activity as measured in a rabbit reticulocyte lye
`sate translation assay. Conjugates with more than two ger
`lonin molecules had a decreased avidity for the antigen,
`which resulted in a less potent molecule. The loss of avidity
`may be due to stearic interference with the antigen—binding
`site or to instability of the over-conjugated IT.
`The humanized M195-gelonin construct had an IDsu of
`15 pM, which is approximately 4500 times more potent
`than free gelonin. When compared to free gelonin,
`the
`humanized immunotoxin appeared more potent
`than the
`murine immunotoxin; this may be due to the significantly
`higher affinity of HuM195 than murine M195 for CD33
`[8, 12, 13]. Neither construct was toxic towards CD3}
`negative cell
`lines at
`the highest concentrations tested.
`Specificity analysis and competition assays showed selec-
`tive binding and activity through CD33 antigen. HL60
`clonogenic assays showed that a reduction by more than a
`factor of 1000 in colony formation was possible with a
`single treatment of IT.
`Murine antibodies usually elicit a human anti-(mouse
`antibody) response, which results in the!rapid clearance of
`murine mAh from serum and loss of therapeutic efficacy
`[10, 20, 34, 35]. The genetically engineered humanized
`antibody HtiM195 appears to be largely nonimmunogenic
`[9]. In addition, the smaller—sized gelonin may be less imr
`munogenic than the
`larger dualrchained ribosome—in-
`activating proteins. Therefore, when conjugated to a non—
`immunogenic humanized mAb a i'eduction in IT im~
`munogenicity is possible, although clinical
`trials will be
`needed to answer this question directly.
`Gelonin has several advantages for use in immuno-
`therapy over other ribosome-inactivating proteins including
`
`IMMUNOGEN 2142, pg. 7
`Phigenix v. Immunogen
`|PR2014-00676
`
`t...) 5.!l-J
`
`Incorporation
`PercentInhibitionofThymidine
`
`J?-01I:D
`
`In}G
`
`20
`
`04
`
`105
`
`10'5
`
`HL60 Concentration (cells/ml)
`
`Fig. 7. Effect of HL60 concentration upon efficacy of HuMlQS—IT.
`HL60 cells at a filial concentration of 3.l25X10475XlU-"’ cells/ml
`were incubated for 5 days at 37 DC with or without IT at a filial
`concentration of 2 [lg/ml. DNA synthesis was determined by 5 h
`incubation with tritiated thymidine. Percentage inhibition was deter
`mined in comparison to control wells without
`IT. The data are
`representative of three experiments
`
`T
`
`,
`..
`
`200000
`
`150000
`
`100000
`
`50000
`
`0
`1E-14 1E-13 1E-12 1E-11
`
`1E-10 1E-03 1E-DB 1E-OT 1E-06
`
`Immunotoxin Concentration (M)
`
`E2
`
`,
`I:
`.2
`I;
`'5
`E-
`
`OU E
`
`d)
`.E
`
`EE
`
`>\
`.C
`p.
`‘U
`2
`.g
`.:
`
`La;
`
`_.11.3
`
`Fig. 8. Cytotoxicity of l—luM l 95~IT on HL60 in the presence of excess
`irradiated bone marrow.
`l—lL60 cells at a final concentration of
`6.67X104 cells/ml
`in the presence (A) or absence (0) of normal
`in‘adiated bone marrow at 1 ><10iJ cells/ml were incubated for 6 days at
`37 0C in the presence of HuGl M19571T at a final concentration of
`100 pg/ml— 10 ug/ml. Levels of DNA synthesis were determined by
`measuring 5 h tritiated thymidine incorporation. The data are repre
`sentative of two experiments
`
`remove non—bound immunotoxin. The cells were resus—
`pended in fresh media and return