Clinical Applications
`of Immunotoxins
`
`Edited by AB. Frankel
`
`With 16 Figures and 11 Tables
`
`*2
`;=@;
`
`
`
`Springer
`
`IMMUNOGEN 2017, pg.1
`Phigenix v, Immu
`en
`|PR2014-
`76
`
`IMMUNOGEN 2017, pg. 1
`Phigenix v, Immunogen
`IPR2014-00676
`
`

`
`Arthur E. FRANKEL, M.D.
`
`Associate Professor of Medicine
`Medical University of South Carolina
`Hollings Cancer Center
`l7l Ashley Avenue
`Charleston. SC 29425-2850
`USA
`
`Cover I//ii.s'rr(iIi0n.' Pietz/red here are molecular mode/.s' of rargetecl
`t0.\‘in.s'. The ribbons are (I/[7/if! carbon bae/(l70ne.s‘ of the following
`targeted t0.\‘ins in eliniecil trials: a) zipper left is D/lB389IL2; /7)
`lower left is LMB-7; c) right is anti CD19-/2R. Green repre.s'ent.s'
`binding domains. Ye//ow rep1'e.s'enr.s' en:_vInc1fie domains. Red is
`/)/0e/\'ec/ Riein B e/Iain.
`
`Cover Design.‘ design & produetian GmbH, Heidelberg
`ISSN O070—2l7X
`ISBN 3-540-64097-5 Springer-Verlag Berlin Heidelberg New York
`
`This work is subject to copyright, All rights are reserved, whether the whole or part of
`the material
`is concerned. specifically the rights of translation. reprinting, reuse of
`illustrations, recitation. broadcasting. reproduction on microfilm or in any other way,
`and storage in data banks. Duplication 0!‘ this publication or parts thereof is permitted
`only under the provisions of the German Copyright Law of September 9, 1965.
`in its
`current version, and permission for use must ‘always be obtained from Springer-Verlag.
`Violations ure liuble for prosecution under the German Copyright Law.
`it‘ Springer-Verlag Berlin Heidelberg 1998
`Library of Congress Catalog Card Number 1512910
`Printed in Germany
`in this
`trademarks, etc.
`registered names.
`The use of general descriptive names.
`publication does not imply. even in the absence ofa specific statement. that such names
`are exempt from the relevant protective laws and regulations and therefore free for
`general use.
`Product
`liability: The publishers cannot guarantee the accuracy of any information
`about dosage and application contained in this book. In every individual case the user
`must check such int'ormation by consulting other relevant literature.
`Typesetting: Scientific Publishing Services (P) Ltd. Madras
`SPIN:
`l0498l()8
`27/3020 — 5 4 3 2 l 0 — Printed on acid—t'ree paper
`
`IMMUNOGEN 2017, pg. 2
`Phigenix v, Immunogen
`|PR2014-00676
`
`#-
`
`IMMUNOGEN 2017, pg. 2
`Phigenix v, Immunogen
`IPR2014-00676
`
`

`
`The Emerging Role of Ricin A-Chain Immunotoxins
`in Leukemia and Lymphoma
`
`A. ENGERT,l E.A. SAUSVlLLE,2 and E. Vn'E1'r/x3
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`Introduction .
`1
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`Preclinical Studies with Ricin A-Chain lmmunotoxins .
`2
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`2.1 First Generation Ricin A-Chain lmmunotoxins
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`2.2 Second Generation lmmunotoxins
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`2.3 Recombinant Ricin A-Chitin lmmunotoxins .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`2.4 Cytotoxic Properties of Ricin A-Chain lmmunotoxin In Vitro .
`2.5 Tissue Distribution and Toxicology of Ricin A-Chain lmmunotoxins in Rodents .
`2.6 The Performance of lmmunotoxins in Animals .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`2.7 lmmunotoxin Cocktails .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`Clinical Trials with Ricin A—Chztin lmmunotoxins .
`3
`.
`.
`.
`.
`.
`.
`.
`3.1 Anti—CD5-Ricin A-Chain .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`3.2 Anti—breast Cancer lmmunotoxin 260F9—rRTA .
`.
`3.3 Anti—B Cell Immunotoxins RFB4—dgRTA and HD 37-dgRTA .
`3.4 Anti-CD25-dgRTA .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`3.5 Vascular Leak Syndrome .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`3.6 Human Anti-immunotoxin Antibodies .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`
`13
`14
`14
`15
`16
`17
`18
`18
`20
`
`20
`20
`21
`21
`24
`24
`25
`
`25
`25
`26
`27
`
`
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`Perspectives
`4
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`4.1 New lmmunotoxins .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`4.2 Vascular Targeting
`.
`4.3 Treatment of Minimal Residual Disease .
`
`Summary
`5
`References
`.
`
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`.
`.
`
`27
`28
`
`1 Introduction
`
`Although some malignancies can be cured by conventional modalities including
`surgery, radiotherapy and polychemotherapy, many cancer patients still ultimately
`die of their disease. The major reason for this poor outcome is the persistence or
`selection of cells which are refractory to conventional treatment. These cells might
`
`' Depzirtment l for Internal Medicine, University of Cologne, 50924 Cologne, Germany,
`e-mail: a.engertm uni-koeln.de
`3Developmental Therapeutics Program. National Cancer Institute. Bethesda. MD 20892, USA
`3Cancer Immunobiology Center and Department of Microbiology, The University of Texas. South-
`western Medical Center. Dallas. TX 75235, USA
`
`IMMUNOGEN 2017, pg. 3
`Phigenix v, Immunogen
`|PR2014-00676
`
`
`
`IMMUNOGEN 2017, pg. 3
`Phigenix v, Immunogen
`IPR2014-00676
`
`

`
`14
`
`A. Engert et al.
`
`be eradicated by new immunotherapeutic agents with different modes of action. In
`this regard, monoclonal antibodies (Moabs) have become available in limitless
`quantities making more selective immunotherapy feasible. These Moabs can bind
`to well-defined antigens expressed on the surface of malignant cells. Unfortunately,
`many Moabs have very weak or no anti-tumor activity when used in their native
`form. Thus, Moabs are often linked to radioisotopes or toxins to increase their
`toxicity. Immunotoxins (ITS) are constructed by chemically or genetically linking
`the antibody moiety to a potent bacterial or plant toxin. To date, the most widely
`used toxin is ricin, which is derived from the seeds of Ric'inu.s' com/mmis (Castor
`bean).
`In this chapter, we describe the preparation of ricin A-chain (RTA)-containing
`ITs and summarize recent data on their preclinical and clinical use.
`
`2 Preclinical Studies with Ricin A-Chain Immunotoxins
`
`2.1 First Generation Ricin A—Chain Immunotoxins
`
`To prepare effective ITS for in vivo therapy, the ligand and toxin must be coupled in
`such a way as to remain stable in both the bloodstream and the tissues and yet be
`labile within the target cell so that the toxic portion can be released into the cytosol.
`RTA-containing ITs have been prepared almost exclusively by chemical cross-
`linking because it
`is diflicult to genetically engineer an IT with a disulfide bond
`between the antibody and the toxin (O’H/me et al. 1990; SPOONER et al. 1994).
`Linkers used to couple ligands to RTA must introduce a disulfide bond be-
`tween the ligand and the RTA. Table 1 summarizes the linkers that have been used
`to accomplish this. Bonds that cannot be reduced (e.g., thioether bonds) render
`these ITs either much less toxic or nontoxic (IVIASUHO et al. 1982), suggesting that
`the RTA is released from the ligand by intracellular reduction. The linker used to
`couple the RTA to the ligand is usually a heterobifunctional cross—linker which
`introduces an activated thiol group into the ligand. When the derivatized ligand is
`mixed with the reduced RTA. the free thiol group in the RTA (its former site of
`attachment to the RTB) displaces the leaving group from the activated thiol group
`introduced into the ligand and forms the disulfide linkage. Since only one or two
`activated thiol groups are usually introduced into the ligand (WAWRZYNCZAK and
`THORPE 1987a; GROS et al. I985), the resultant IT is contaminated with both free
`RTA and free ligand. The free RTA, much of the unreacted antibody, and high
`molecular weight aggregates are removed by gel permeation chromatography
`(CUMBER et al. 1985). The remaining free antibody is removed by adsorbing the IT
`by its RTA moiety to columns of either blue Sepharose (KNOWLES and THORPE
`1987) or immobilized anti~RTA antibodies (VITETTA et al. I987), washing away the
`free antibody and eluting the RTA-containing IT (IT—A) with high salt or low
`pH. It should be noted that while contamination with RTA is not usually a
`
`
`
`IMMUNOGEN 2017, pg. 4
`Phigenix v, Immunogen
`|PR2014-00676
`
`IMMUNOGEN 2017, pg. 4
`Phigenix v, Immunogen
`IPR2014-00676
`
`

`
`The Emerging Role of Ricin A-Chain lmmunotoxins
`
`15
`
`Table 1. Linkers used to prepare rlcin A-chaiwconlaining
`immunotoxins
`
`Cross-Linker+
`
`IT Structure
`
`None
`SATA
`spor
`2-IT
`
`SMPT
`
`Fab—S S— Q)
`lgG—N1{—CO—CH2—SS~®
`lgG—NI-I—CO—CH2—C1—I2~SS—@
`+I\IIH 2
`1gG—NI-I—C—C1I2—CII2—CII —ss—®
`CH3
`1gG—NH—co—@— dH—ss—@
`
`“From Wawrzynczak and Thorpe (Wawrzynczal< and
`Thorpe 1987).
`+ 2- I T = 2—iminothiolane; SATA = N-succinimidyl—
`S-thioacetate; SPDP. N—succinimidyl-3-(2-pyriclyldihio)
`propionate;
`SMPT = N—succinimidyl-oxycarbonyl-a-methy|—a(2~pyr0—
`dyldithio) tuluene.
`
`problem in vivo, contamination with free antibody can present a significant
`problem (BLAKEY and THORPE 1988), since the half-life of an intact antibody is
`much longer than that of an IT (BLAKEY et al. 1987; BOURRIE et al. 1986; BYERS
`et al. 1987). Thus, after a period of time in vivo, the free antibody is present in the
`blood and tissues in greater amounts than the IT and can compete for binding sites
`on the target cells, effectively reducing the potency of the IT.
`
`2.2 Second Generation Immunotoxins
`
`The “first generation" heterobifunctional cross-linkers, such as SPDP or 2—IT,
`generated disulfide bonds that were quite unstable in vivo, releasing RTA and
`antibody with a T”; of 6-811 (FULTON et al. 198821; BLAKEY et al. 1987; THORPE et al.
`1987a). Breakdown was probably due to reduction by glutathione, albumin, and
`other thiol-containing molecules in the bloodstream and tissues. This problem has
`been solved by the synthesis of new cross-linkers, e.g., SMPT (THORPE et al. l987b),
`which introduce hindered disulfide bonds having phenyl and/or methyl groups onto
`carbon atoms adjacent to the disulfide bond. ITs prepared with these hindered
`cross-linkers are much more stable (T./3, approximately 2 days) (THORPE et al.
`1987b) and have improved anti-tumor activity (THORPE et al. 1988) compared to
`their predecessors.
`In addition to utilizing better cross—linkers, some second generation ITs were
`also smaller. Since Fab and some Fab’ fragments of antibodies have a free cysteine
`residue near the hinge region (STANWORTH and TURNER 1978), these can be used to
`form a disulfide bridge with the free cysteine residue of the RTA. The coupling is
`generally accomplished by activating the thiol group in the Fab/Fab’ fragment with
`El1man’s reagent (to provide a good leaving group) and mixing the derivatized Fab/
`
`;
`
`1
`
`5
`
`‘
`
`
`
`_—_r.,._..._.,.......,,
`
`l
`
`I
`
`
`
`IMMUNOGEN 2017, pg. 5
`Phigenix v, Immunogen
`|PR2014-00676
`
`IMMUNOGEN 2017, pg. 5
`Phigenix v, Immunogen
`IPR2014-00676
`
`

`
`I6
`
`A. Engert et al.
`
`Fab’ with the reduced RTA (R/\so and GRu~‘r~‘iN 1980). The Fab/Fab’—RTA then
`forms by a thiol-disulfide exchange reaction. Since the thiol group of the Fab/Fab’
`fragment
`is distant from its antigen-combining site (STANWORTH and TURNER
`I978), and since there is usually only one free thiol group on the fragments derived
`from most subclasses ofmurine Moabs (PARHAM 1983; SVASTI and MILSTEIN I972),
`this assures the attachment of one RTA per Fab/Fab’ fragment at a site distant
`from the combining site. Unlike the IgG-RTA, such constructs, therefore, retain
`full antigen—binding activity.
`fucose-terminating
`and
`and contains mannose
`RTA is
`glycosylated
`oligosaccharides (KIMURA et al.
`I988). These sugars are recognized by avid re-
`ceptors on both the parenchymal and nonparenchymal cells of the liver and by cells
`ofthe reticuloendothelial system (RES) (BLAKEY et al. 1988; SKILLETER et al. 1981).
`Hence, IT—RTAs prepared from native RTA are cleared rapidly from the blood-
`stream. and this may have been at least partially responsible for the failure of first
`generation IT—RTAs to reach their target cells in vivo and elicit their intended anti-
`tumor effects. In addition. entrapment of the IT-RTAS by the liver results in liver
`damage (JANSEN et al. 1982). although this is fairly modest because IT-RTAS taken
`up through the carbohydrate recognition pathways are probably routed to the
`lysosomes and destroyed.
`i.e.,
`The problem of liver recognition of RTA was solved by destroying,
`“deglycosylating” the mannose and fucose residues on the ricin molecule by a
`simple procedure involving periodate oxidation and cyanoborohydride reduction at
`low pH (THORPE et al. 1985b). This procedure does not afi"ect the enzymatic activity
`of RTA or its ability to function as an IT (BLAKEY et al. 1987). ITs constructed with
`deglycosylated RTA (dgRTA) prepared with intact antibodies and stable linkers
`arc long—lived in mice lacking target cells for which the IT is specific. The IT is
`cleared only about twice as quickly as is native mouse IgG (TI-IORPE et al. 1987a,
`THORPE ct al. 1988).
`
`2.3 Recombinant Ricin A-Chain Immunotoxins
`
`The CDNA encoding the precursor of ricin (preproricin) has been cloned (LAMB
`et al. 1985) and the segment of DNA encoding the A~chain has been expressed in
`E.s‘c'/zeric/ziu co/1' (O’HARE et al. 1987). The expressed RTA was devoid of carbo~
`hydrate and did not
`localize in the liver. A technique was also developed for
`synthesizing cytotoxic RTA fusion proteins in E. coli (O’HARE et al. 1990). The
`gene for the ligand (staphylococcal protein-A) was ligated to the DNA encoding
`RTA via an intervening spacer sequence derived from the diphtheria toxin gene.
`The single chain fusion protein was expressed in E. coli. The function of the
`diphtheria toxin spacer was to form a disulfide—bonded loop between the ligand and
`the A—chain which could be subsequently nicked with trypsin to yield a two chain
`structure in which the ligand and the A—chain were joined by a disulfide bond. The
`nicked fusion protein was highly cytotoxic to Ig—coated target cells in vitro. An
`active IT prepared with an anti-breast cancer Moab biochemically cross-linked to
`
`IMMUNOGEN 2017, pg. 6
`Phigenix v, Immunogen
`|PR2014-00676
`
`IMMUNOGEN 2017, pg. 6
`Phigenix v, Immunogen
`IPR2014-00676
`
`

`
`
`
`._.._—...._.......,_.r-.....,,__,..._,..,...2,7
`
`The Emerging Role of Ricin A—Chain lmmunotoxins
`
`17
`
`recombinant RTA has been prepared and tested in humans (BJORN et al. 1985;
`FRANKEL et al. 1985; GOULD et 211. 1989).
`
`2.4 Cytotoxic Properties of Ricin A—Chain Immunotoxin In Vitro
`
`RTA-containing 1Ts Show virtually complete specificity in their cytotoxic effect on
`target cells in vitro bec21use their only means of binding to cells is through the ligand
`portion of the IT. Fc binding, ifit occurs, appears not to route the IT to an intra-
`cellular compartment favorable for RTA release and cell killing. The major disad-
`vantage ofIT-RTAS, however, is that they have variable cytotoxic potencies (BJORN
`et al. 1985); only about 25% of such ITS are potently toxic to their target cells.
`It is now generally accepted that a variety of factors concerning Moabs or
`ligands play major roles in determining whether or not they will make elfective ITS
`(GHETIE ct al. 1988; RAMAKRISHNAN and HOUSTON 198421; SHEN et al. 1988;
`ENGERT et al. 199021). These variables include specificity, affinity, internalization
`and intracellular routing. For example,
`those antibodies or ligands with low
`binding atfinity generally make poor ITS while those with high affinity are usually
`quite good. This is probably the main reason why Fab’—RTAS having only one
`antigen-combining site tend to be threefold to several 100-fold less effective than
`their intact antibody counterparts, which have higher avidity because of their
`bivalency. Secondly, the cell surface molecule that the antibody recognizes plays a
`key role in determining the efficacy of the IT (PRESS et al. 1986, 1988; MAY et al.
`1990). In general, cell surface molecules such as growth factor receptors, which
`continuously recycle into endosomal compartments or which can be induced to do
`so by binding to an antibody or other ligand, make effective targets for ITS. Cell
`Surface molecules which are not readily internalized or which are internalized into
`intracellular compartments unfavorable for RTA translocation (e.g.,
`the lyso~
`somes) make poor ITS no matter how high their avidity. It has also been reported
`that the target antigen, which is often a large cell surface glycoprotein, can have
`numerous epitopes and that antibodies against different epitopes can differ in their
`eiTective11ess21S ITS (PRESS et al. 1988; MAY et al. 1990). The most likely explanation
`for this finding is that Moabs may have to recognize epitopes close to the plasma
`membrane for the RTA to insert into the membrane and be translocated across the
`
`membrane into the cytosol after endocytosis.
`RTA-containing ITS that are weakly effective often Show greatly improved
`toxicity in the presence of lysosomotropic amines (e.g., ammonium chloride,
`chloroquine) (CASELLAS et al. 1984; RAMAKRISHNAN and HOUSTON 1984), car-
`boxylic ionophores (e.g., monensin) (RASO and LAWRENCE 1984), RTB chains
`(MC1N'rosH et 211. 1983; YouLE and NEVILLE 1982) and Moab~RTB chains (VITETTA
`et al. 1983, 1984). Two- to 1,000-fold improvements in IT potency are common.
`The lysosomotropic amines and the carboxylic ionophores probably work by
`slowing the fusion of endosomes with lysosomes (R/\So 1988) (where RTA—chains
`are enzymatically destroyed) or by delaying the transit of the IT through a com-
`partment favorable for RTA translocation, for example, the tranSGolgi. The RTB
`
`
`
`IMMUNOGEN 2017, pg. 7
`Phigenix v, Immunogen
`|PR2014-00676
`
`IMMUNOGEN 2017, pg. 7
`Phigenix v, Immunogen
`IPR2014-00676
`
`

`
`18
`
`A. Engert et al.
`
`chains and Moab-RTB chains probably work by protecting the RTA from de-
`gradation O1‘ enhancing their translocation across the membranes of intracellular
`compartments. Lysosomotropic amines also potentiate the nonspecific activity of
`any contaminating toxin or RTB (FULTON et al. 1986). Exquisite care must,
`therefore, be taken to eliminate RTB Or toxin from a purified IT or enhancement
`may result from contamination rather than from a direct elTeet on the IT.
`An assay has been developed to screen Moabs in order to predict which ones
`will make effective ITS (TILL et al. 1988). This assay involves binding the Moab (in
`the form of a tissue culture supernatant, ascitcs, or purified antibody) to the target
`cells and then treating the coated cells with a secondary Fab’ or Fab fragment of an
`anti—mouse, rat, or human immunoglobulin linked to an RTA (TILL et al. 1988). In
`most cases, the degree of killing predicts quite accurately the potency of that Moab
`when it is coupled directly to the RTA-chain. This assay is an important devel-
`opment since previously each Moab had to undergo time-consuming purification,
`linkage to the RTA-chain, and evaluation by cytotoxicity assays on target cells.
`
`2.5 Tissue Distribution and Toxicology of Ricin A-Chain Immunotoxins
`in Rodents
`
`Of all the ITS so far studied in depth, only those prepared with dgRTA appear not
`to cause liver damage in rodents. At doses of IT—dgRTAs approaching the LD50,
`the only normal tissue damage visible in mice and primates is in the crypts of
`Lieberkuhn, the small intestine (JANSEN et al. 1982), the proximal tubule of the
`kidney (J/\NsEN et al. 1982), and skeletal muscle (FULTON et al. 1988). In vitro,
`dgRTA/RTA can damage human endothclial cells (SOLER-RODRIGUEZ et al. 1993).
`None of these effects appear to be severe enough to account for death, and the cause
`of death at high doses of IT-As is currently known. The Fab’—dgRTAs induce
`similar toxic side effects except that,
`in some experiments, the myositis is more
`marked, possibly because Fab'—dgRTAs permeate extravascular
`tissues more
`readily because of their smaller size. By contrast, the LD50 of Fab’-dgRTAs is three-
`to fivefold greater (On a total protein basis) than that of IgG-dgRTAs (FULTON et
`al. 1988), probably because they are more rapidly cleared (FULTON et al. 1988).
`
`2.6 The Performance of Immunotoxins in Animals
`
`Animal models have facilitated the evaluation and refinement of different types of
`IT constructs. First—generation IT—As containing native RTA and unstable linkers
`were developed into second-generation IT-As. These contain dgRTA and stable
`linkers and have dramatically superior anti-tumor activity as compared with their
`predecessors (VITETTA et al. 1987; THORP13 et al. 1988). Moreover, Fab’—dgRTAs
`have been compared with IgG—dgRTAs and, depending on the test system, have
`shown weaker or equivalent anti-tumor activity, but
`lower
`immunogenicity
`(FULTON et al. 1988).
`
`IMMUNOGEN 2017, pg. 8
`Phigenix v, Immunogen
`|PR2014-00676
`
`IMMUNOGEN 2017, pg. 8
`Phigenix v, Immunogen
`IPR2014-00676
`
`

`
`Table 2. Curative ellect of dillerent inimuiiotoxins in SCID mice with disseminated human tumors
`
`The Emerging Role of Ricin A-Chain lmmuiiotoxiiis
`
`19
`
`I-l uina ii
`xenogra ft
`
`Tum or
`type
`
`lnoculati on
`
`1 in munotoxin
`construct
`
`Treatment
`Speci licity
`
`Time a liter
`tumor
`inoculation
`
`Tlierapeutic"
`elllect
`(%CR)
`
`L540Cy
`
`Daudi
`
`Daudi
`
`Hodgkin's
`lyniphoma
`
`Bui'1<itt’s
`lymphoma
`
`Burkitt‘s
`lymphoina
`
`i.v.
`
`i.v.
`
`i.v.
`
`CD25
`70KD
`
`CD22/
`CD19
`
`CD22/
`CD19
`
`RFT5-dg RTA
`lRac—c|g RTA
`
`RFB4-dgRTA +
`HD37(Fab)3
`
`RFB4-dgRTA +
`HD37-dgRTA +
`Doxorubicin
`(cytoxaii)
`(campothecin)
`
`1 day
`12 days
`1 day
`1 day
`
`1 day
`
`95
`46
`93
`100
`
`100
`
`“ Percent of mice in complete remission: CR: longterm, tumor—free survival
`
`It has been established in animal models that marked anti-tumor effects,
`
`prolonged remissions, and even cures (THORPE et al. 1985a, 1988; JANSEN et al.
`1980) can be achieved with doses of RTA—lTs that are well tolerated (Table 2). In
`general, tumors that are readily accessible to the bloodstream appear to be most
`responsive. In addition, the effectiveness of combinatorial therapy has been well
`documented in animal models (GHETIE et al, 1996). For example, access of ITS to
`solid tumors and anti~tu1nor activity can be improved by coadministration of ad-
`renergic blockers, which may act by selectively constricting normal vasculature and
`increasing the tuinor-to—normal—tissue perfusion ratio (SMYTH et al. 1987). Coad—
`ministration of the carboxylic ionophore monensin, either free (RAMAKRISHNAN
`et al. 1989) or conjugated to serum albumin (CASELLAS and .l/\NSEN 1988), can also
`potentiate the anti-tumor activity of an lT—A. It will probably be necessary to give
`ITS in combination with conventional cheinotherapeutic drugs or radiotherapy
`because the modes of action of the difi"erent therapies do not overlap; thus, the
`resistant mutants that escape one type of therapy succumb to the other. In this
`regard, results in SCID mice xenografted with human tumors demonstrate that ITS
`work well in combination with chemotherapy (GHETIE et al. 1994, 1996).
`The administration of ITS to animals with intact immune systems leads to an
`antibody response against both components, precluding repeated treatments. lin-
`inunosuppressive regimens can delay anti-IT responses, but may also compromise
`the host’s ability to suppress the growth of residual tumor. One way to decrease the
`iininuiiogeiiicity of ITS is to modify them with monoinethoxy-polyethyleneglycol
`(PEG). PEG has been used to decrease the iminunogenicity and increase the blood
`residence time of various enzymes (KATRE 1993) and Igs (KITAMURA et al. 1991,
`1996). PEG—RTA has been conjugated to an anti—breast cancer antibody but the
`resulting IT showed a tenfold decrease in its ability to inhibit protein synthesis in a
`cell-free system. Unexpectedly, its cytotoxicity in vitro against a breast cancer cell
`line was unallected. Because RTA has only two lysine residues (susceptible to
`derivatization with PEG), the number of PEG molecules per RTA may not be
`
`
`
`IMMUNOGEN 2017, pg. 9
`Phigenix v, Immunogen
`|PR2014-00676
`
`IMMUNOGEN 2017, pg. 9
`Phigenix v, Immunogen
`IPR2014-00676
`
`

`
`20
`
`A. Engert et al.
`
`sufficient to confer a longer half—life and lower immunogenicity (V. GHETIE, un-
`published results).
`immunosuppressives agents which block
`More recent studies involve novel
`either T/B cell stimulation or eostimulation (LAM/\N et al. I996; LENSCHOW et al.
`1996; SIEGALL et al. 1997) and preliminary results are encouraging (E. NOON and E.
`S. VITIETTA, submitted to Clin, Cancer Res. The effect of immunosuppresive agents
`on the immunogenicity and eificacy of an Immunotoxin in mice). However‘, until
`specific tolerance to an IT can be induced, it is probably advisable to administer the
`IT in a short course prior to the development of a primary antibody response.
`
`2.7 Immunotoxin Cocktails
`
`Another problem in using Moab~based anti-tumor reagents is the selection of anti-
`gen—deficient mutants causing late relapses in experimental animals (THORPE et al.
`I988). A solution would be to use IT cocktails, i.e.. combinations of ITs directed
`against different antigens or epitopes on the same target cell. Superior effects of IT
`Cocktails have been demonstrated for both, non-Hodgkin’s and Hodgkin’s lym-
`
`phoma (NI-IL) (GHETIE et al. 1992; ENGERT et al. 1995). Lymphomas are well suited
`for the use oflT cocktails since several potent ITS against different target antigens are
`available. The superiority of the IT cocktails can be explained by the fact that most
`tumors contain a proportion ofcells which express the target antigen at low density at
`the time of treatment. These cells might be a “dormant” subpopulation having re-
`duced numbers of activation markers which are nevertheless capable of fully reen-
`
`tering proliferation (PANTEL et al. 1993). Antigen-deficient cells which survive after
`treatment with a single IT were killed when that IT was combined with a second or
`third IT directed against a different antigen on the same cell.
`IT cocktails consisting of ITs against CD25, CD30, and IRac (70kDa) on L540
`Hodgl<in‘s cells in any combination have demonstrated superior anti-tumor activity
`to single ITs, both in vitro and when used to treat solid human Hodgkin’s tumors in
`nude mice (ENGERT et al. 1995). In SCID mice bearing disseminated Daudi lym-
`phomas, an IT cocktail consisting of I-ID37 (anti—CDl9)-dgA and RFB4 (anti-
`CD22)-dgA was shown to kill an equivalent of 5 logs of tumor cells, while HD37-
`dgRTA alone killed 2 logs and RFB4-dgRTA alone killed 4 logs (GHETIE et al.
`1992). Studies in nude mice have reported superior elTects of IT cocktails against
`non-T leukemia cells with ITS against CALLA (SN5-A) and a leukemia—associated
`surface glycoprotein (SN6-A) (I-IARA et al.
`l988).
`
`3 Clinical Trials with Ricin A-Chain Immunotoxins
`
`3.1 Anti-CD5-Ricin A-Chain
`
`Initial studies with RTA-based ITs in non—Hodgkin‘s lymphoma examined RTA
`
`conjugated to the Moab anti-CD5. This construct produced four partial responses
`
`IMMUNOGEN 2017, pg. 10
`Phigenix v, Immunogen
`|PR2014-00676
`
`IMMUNOGEN 2017, pg. 10
`Phigenix v, Immunogen
`IPR2014-00676
`
`

`
`The Emerging Role of Ricin A-Chain Immunotoxins
`
`21
`
`lymphoma (LEMAISTRE et al. 1991). The
`in 14 patients with cutaneous T cell
`dominant toxicity was the occurrence of vascular leak syndrome (VLS) and re-
`versible hepatic dysfunction. The latter was most likely due to the localization of
`the IT in the liver, since the RTA was fully glycosylated.
`
`3.2 Anti—breast Cancer Immunotoxin 260179-rRTA
`
`When recombinant RTA was conjugated to the anti-breast carcinoma cell antibody
`260139 and used in patients with advanced breast cancer, VLS dominated the
`toxicity profile (WEINER et al. 1989). In addition, a neurologic syndrome consisting
`of sensorimotor neuropathy complicated the use of this agent. This unexpected side
`effect was ultimately linked to an unsuspected cross—reaction of the antibody with a
`determinant on either Schwann cells or myelin (GOULD et al. 1989).
`
`3.3 Anti—B Cell Immunotoxins RFB4—dgRTA and HD 37-dgRTA
`
`The CD22 and CD19 determinants, expressed on adult B cell lymphomas, have
`been targeted in a series of clinical studies using dgRTA ITs. The Fab’ fragment of
`the CD22-directed antibody RFB4 was linked to dgRTA and the resulting IT was
`studied in a regimen involving bolus doses administered every 48h for up to six
`doses. Partial remissions (PR3) of 1-4 months duration were noted in 38% of 15
`patients, and VLS was the dominant reversible toxicity. Expressive aphasia and
`rhabdomyolysis were close limiting toxicities. The Tug of the IT was short, as can be
`expected for a relatively small molecule (VITETT/\ et al. 1991). A similarly designed
`phase-I trial using bolus administration of the RFB4—dgRTA in its IgG form gave
`five PR5 and one complete remission (CR) lasting 30~78 days in 26 patients. VLS
`was the dose-limiting toxicity (AMLOT et al. 1993). In addition, there was a trend
`toward decreased toxicity in patients with evidence of bulky disease including
`splenomegaly. The most prominent clinical trials in lymphoma patients are suin-
`marized in Table 3.
`In an effort to reduce the incidence of VLS, a phase-I trial using a continuous
`infusion regimen was conducted (SAUSVILLE et al. 1995). In this trial, the IT was
`administered continuously over 19211 (8 days). Despite this, there was essentially no
`difference in the maximal tolerated dose (MTD; 19mg/mg per 8 days) compared to
`the intermittent bolus dosing schedule. This study therefore concluded that there
`was no improvement in therapeutic index by administering the IT as a continuous
`infusion. The RFB4—dgRTA construct showed evidence of clinical activity, with
`four of 16 evaluable patients achieving PRS. The highest serum concentrations
`(> l00Ong/ml) on days 2 or 3 of the infusion were predictive of severe VLS. This
`observation was the first clear evidence of a relationship between the IT concen-
`tration in the blood and the occurrence of VLS. The T1,; of the IT was also
`variable, ranging as high as 23h in certain patients. Small numbers of circulating
`tumor cells, detectable in some cases only by flow cytometric analyses, correlated
`
`
`
`IMMUNOGEN 2017, pg. 11
`Phigenix v, Immunogen
`|PR2014-00676
`
`IMMUNOGEN 2017, pg. 11
`Phigenix v, Immunogen
`IPR2014-00676
`
`

`
`22
`
`A. Engert et al.
`
`
`
`
`
`mfiacamEo~EE.3EwEx9om=EE_Em:o-<Eu:~3323.:786:0.m28...“.
`
`
`
`
`
`
`
`
`
`39._wa<EE>Eis:4<32:1;.§m:...€..3>cemee;V<§w3muE,n_£NRC,_:z.m
`
`
`
`
`
`<m<mEzoiofimfi73;ad
`
`
`
`82.23E§<E9%JG3;0.;<2<zcm;.2935,3>memesnw<Ewu.E2m.E$_SouEz.m
`
`
`
`
`
`
`
`<.m<Emabo.¢:omn:#:>1m.mL>mQ
`
`
`
`22gm3B:_>S<mESe25<32:23mi,zoaeemamsaou<E%-E2m-XEmSou4:72
`
`
`
`
`
`
`
`R2.33mzopmEmm;inmm;as<32:m3.2233.3>comesaV<Emc.:om-ow_EQU.52.».
`
`32:ml55
`
`m_wo:m.Go.Sm
`
`32$.m_£%EoEm€TE;25
`
`
`
`
`
`
`
`
`
`32.35mzohmE9;M2<32:3emofibofi.B>ceaseaoaacou<Emc.:a:-om_Sou.Ez.m
`
`
`
`
`
`:2.35:28EW:<m<E<2<m5$35.3>memes3Vflame.:.o:.owEmug:z-m
`
`$2:mi$3
`
`
`
`
`
`32.33bmozmyaCa<32:23uzmsa.m1_>comes3V<53-2r.E.ow_mooMEWS:
`
`
`
`wfibozfiownnfim:m.mI_>5
`
`<m<mBKawieTE;E5E295.
`
`
`
`
`
`
`
`ouzmhdumomcommom.€uEmmo::EE_E_oc,.oH:o:3:nE<Exo._.:um::<umawmfla
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`”:o_mmE._Ev.E_QEoodoumonuonucmEucéam:mE::.<m<ImozuoncmmomaofiécmSEE:.<2<IUEo€E$V32§=umm>.w»_>:..Eo:9E:m,:EmvoI.:o:{EZ
`
`
`
`
`
`
`
`
`
`
`
`
`
`.commmE_u._3:3ME
`
`IMMUNOGEN 2017, pg. 12
`Phigenix v, Immunogen
`|PR2014-00676
`
`IMMUNOGEN 2017, pg. 12
`Phigenix v, Immunogen
`IPR2014-00676
`
`

`
`
`
`The Emerging Role of Ricin A-Chain Immunotoxins
`
`23
`
`with decreased toxicity. These patients also tended to have increased volumes of
`distribution (Vd) of the IT with lower circulating IT levels achieved during the
`infusion. A notable trend using both the bolus and continuous infusion regimens
`was the likelihood of a better clinical response in patients with smaller tumor
`burdens (< lO0cm3), reinforcing the potential value of this IT in patients with
`minimal disease.
`
`A major problem in the use of anti—CD22—directed ITs is the variable expres-
`sion of CD22 even on cells within the same tumor, and the fact that at best only
`65%—70°/o of tumors are CD22+. In contrast,
`the CD19 determinant is more
`ubiquitously expressed on cells in adult lymphoid tumors. The anti—CDl9—Moab
`HD37 was therefore used to construct HD37-dgRTA, and this agent was studied in
`regimens involving both intermittent bolus and continuous infusion. Although this
`IT was not as cytotoxic as RFB4—dgRTA, it was still highly potent. When given as
`an intermittent bolus infusion every 48 h for four closes, the MTD was 16mg/mg
`per 8 days. VLS again define