on October 22, 2014
` on October 22, 2014
` on October 22, 2014
` on October 22, 2014
` on October 22, 2014
`
`www.sciencemag.org
`www.sciencemag.org
`www.sciencemag.org
`www.sciencemag.org
`www.sciencemag.org
`
`Downloaded from
`Downloaded from
`Downloaded from
`Downloaded from
`Downloaded from
`
`H. Will, J. Virol. 62, 3832 (1988).
`46. H. Popper et al., Proc. Nail. Acad. Sc. U.SA. 84, 866 (1987).
`47. A. Dejean, L. Bougueleret, K. H. Grzeschik, P. Tiollais, Nature 322, 70 (1986); T.
`Moroy et al., ibid. 324, 276 (1986).
`48. A. F. von Lringhofen, S. Koch, P. H. Hofichneider, R. Koshy, EMBOJ. 4, 249
`(1985); M. Wollersheim, U. Debelka, P. H. Hofichneider, Oncogene 3, 545
`(1988); S. Takada and K. Koike, Proc. Nad. Acad. Sci. U.SA. 87, 5628 (1990).
`49. H. P. Wang and C. E. Rogler, Cytogenet. Cell Genet. 48, 72 (1988).
`50. I. C. Hsu et al., Nature 350, 427 (1990).
`51. I. Saito et al., Proc. Nati. Acad. Sci. U.SA. 87, 6547 (1990).
`52. J. P. Sundberg, Contrib. Oncol. 24, 11 (1987).
`53. E. M. de Villiers, J. Virol. 63, 4898 (1989); personal communication.
`54. G. Orth, in The Papovaviridae, N. P. Salzman and P. M. Howley, Eds. (Plenum,
`New York, 1987), p. 199.
`55. D. M. Parkin, J. Stjernsward, C. S. Muir, Bull. WHO 62, 163 (1984).
`56. H. zur Hausen, Biodhim. Biophys. Acta 417, 25 (1975); Cancer Res. 36, 794 (1976).
`57.
`, Cancer Res. 49, 4677 (1989).
`58. T. Matsukura, S. Koi, M. Sugase, Virology 172, 63 (1989); A. P. Cullen, R. Reid,
`M. Campion, A. T. LUrincz, J. Virol. 65, 606 (1991).
`59. E. Schwarz et al., Nature 314, 111 (1985); C. Yee et al., Am. J. Pathol. 119, 361
`(1985).
`60. K. Munger et al., J. Virol. 63, 4417 (1989); P. Hawley-Nelson et al., EMBOJ.
`8, 3905 (1989); S. Watanabe, T. Kanda, K. Yoshiike, J. Virol. 63, 965 (1989).
`61. V. Band, D. Zajchowski, V. Kulesa, R. Lager, Proc. Nad. Acad. Sci. U.SA. 87,
`463 (1990).
`62. D. J. McCance, R. Kopan, E. Fuchs, L. A. Laimins, ibid. 85, 7169 (1988); J. B.
`Hudson, D. J. Bedell, D. J. McCance, L. A. Laimins, J. Virol. 64, 519 (1990); C.
`D. Woodworth et al., Cancer Res. 50, 3709 (1990).
`63. P. J. Hurlin et al., Proc. Nad. Acad. Sci. U.SA. 88, 570 (1991); G. Pecoraro, M.
`Lee, D. Morgan, V. Defendi, Am. J. Pathol. 138, 1 (1991).
`64. C. P. Crumetal., J. Cell. Biochem. (Suppl.) 9c, 70 (1985); Y. S. Fu, J. W. Reagan,
`R. M. Richart, Gynecol. Oncol. 12, 220 (1981).
`
`65. N. Dyson, P. M. Howley, K. Munger, E. Harlow, Science 243, 934 (1989); B. A.
`Werness, A. J. Levine, P. M. Howley, ibid. 248, 76 (1990).
`66. D. Eliyahu et al., Proc. Nail. Acad. Sci. U.S.A. 86, 8763 (1989); C. A. Finlay, P.
`W. Hinds, A. J. Levine, Cell 57, 1083 (1989).
`67. M. Scheffner et al., Cell 63, 1129 (1990).
`68. L. Pirisi et al., J. Virol. 61,1061 (1987); M. Durst et al., Oncogene 1, 251 (1987);
`P. Kaur and J. M. McDougall, J. Virol. 62, 1917 (1988).
`69. M. Scheffner, K. Munger, J. C. Byrne, P. M. Howley, Proc. Nail. Acad. Sci. U.S.A.
`88, 5523 (1991).
`70. M. von Knebel-Doeberitz, T. Oltersdorf, E. Schwarz, L. Gissmann, Cancer Res.
`48, 3780 (1988); M. von Knebel-Doeberitz et al., unpublished observations; T.
`Crook, J. P. Morgenstern, L. Crawford, L. Banks, EMBOJ. 8, 513 (1989).
`71. M. von Knebel-Doeberitz, T. Bauknecht, D. Bartsch, H. zur Hausen, Proc. Nail.
`Acad. Sci. U.SA. 88, 1411 (1991).
`72. J. A. Di Paolo et al., Oncogene 4, 395 (1989); M. Dirst, D. Gallahan, J. Gilbert,
`J. S. Rhim, Virology 173, 767 (1989).
`73. F. X. Bosch et al., J. Virol. 64, 4743 (1990).
`74. M. Dirft et al., ibid. 65, 796 (1991).
`75. H. zur Hausen, Behring Inst. Mitt. 61, 23 (1977).
`76. L. Braun, M. Dirst, R. Mikumo, P. Grupposo, Cancer Res. 50, 7324 (1990); C.
`D. Woodworth, V. Notario, J. A. Di Paolo, J. Virol. 64, 4767 (1990); S.
`Yasumoto, A. Taniguchi, K. Sohma, ibid. 65, 2000 (1991).
`77. F. Rosi, M. Diirst, H. zur Hausen, EMBOJ. 7, 1321 (1988).
`78. F. Rosi et al., ibid. 10, 1337 (1991).
`79. H. Romanczuk, F. Thierry, P. M. Howley, J. Virol. 64, 2849 (1990).
`80. M. Boshart et al., EMBOJ. 3, 1151 (1984).
`81. F. RbsI, E.-M. Westphal, H. zur Hausen, Mol. Carcinog. 2, 72 (1989).
`82. P. A. Jones et al., Proc. Nad. Acad. Sci. U.SA. 87, 6117 (1990).
`83. G. Madashewski, L. Banks, D. Pim, L. Crawford, Eur. J. Biochem. 154, 665
`(1986).
`84. P. J. Saxon, E. S. Srivatsan, E. J. Stanbridge, EMBOJ. 5, 3461 (1986); M. Koi
`et al., Mol. Carcinogen. 2, 12 (1989).
`
`Recombinant Toxins for Cancer Treatment
`IRA PASTAN AND DAVID FITZGERALD
`
`Rrecombinant toxins target celi surface receptors and
`antigens on tumor cells. They kill by mechanisms differ-
`ent from conventional chemotherapy, so that cross resis-
`tance to conventional chemotherapeutic agents should not be
`a problem. Furthermore, they are not mutagens and should
`
`not induce secondary malignancies or accelerate progression
`ofbenign malignancies. They can be mass-produced cheaply
`in bacteria as homogeneous protins. Either growth factor-
`toxin fusions or antibody-toxin fusions can be chosen, de-
`pending on the cellular target
`
`R MECOMBINANT TOXINS ARE HYBRID CYTOTOXIC PROTEINS
`made by recombinant DNA technology that are designed to
`selectively kill cancer cells. The cell-targeting moiety can be
`a growth factor or a single chain, antigen-binding protein. The toxic
`moiety is a portion of a bacterial or plant toxin. Immunotoxins are
`similar in concept but are composed of antibodies chemically linked
`to toxins.
`More than 30 years ago chemotherapeutic drugs began to be used
`to treat cancer as a supplement to surgery and radiation therapy.
`Now that several decades have passed, it is clear that the current
`generation of chemotherapeutic drugs can achieve cures of certain
`leukemias and lymphomas and, in the adjuvant setting, prolong the
`lives of patients with breast cancer, ovarian cancer, and several other
`types of cancer. Because chemotherapy is not a cure for the common
`types of cancer in adults, new therapies must be developed.
`
`The authors are in the Laboratory of Molecular Biology, Division of Cancer Biology,
`Diagnosis, and Centers, National Cancer Institute, National Institutes of Health,
`Bethesda, MD 20892.
`
`22 NOVEMBER 1991
`
`One approach is to target a cytotoxic agent to the cancer cell (1).
`To accomplish this, the cytotoxic agent is attached to an antibody or
`a growth factor that preferentially binds to cancer cells. The targets
`for this type of therapy can be growth factor receptors, differentia-
`tion antigens, or other less characterized cell surface antigens. It is
`now established that many cancers overproduce growth factor
`receptors that can function as oncogenes and promote the growth of
`the cancer cells (2-4). For example, the epidermal growth factor
`receptor is present in large amounts (up to 3 x 106 receptors per
`cell) in many squamous cell and epidermoid carcinomas, glioblasto-
`mas, and some metastatic ovarian and bladder cancers (5-7). Normal
`cells contain as many as 3 x 105 receptors per cell (8). The
`interleukin-2 (IL-2) receptor is present in substantial numbers on
`the cells of patients with adult T cell leukemia (ATL; 3 x 104
`receptors per cell) and in lower numbers in various other lymphoid
`malignancies (9).
`Differentiation antigens that occur on normal cells such as B
`lymphocytes are often also present on tumor cells such as B cell
`lymphomas. Because such antigens are not present on the stem cells
`
`ARTICLES
`
`1173
`
`IMMUNOGEN 2146, pg. 1
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`that produce B cells, any mature B cells that are killed by targeted
`therapy will be replaced from the stem cell population, whereas the
`cancer cells will not be replaced (10). Finally, there are antigens
`preferentially expressed on cancer cells whose functions are not yet
`understood. Some ofthese, such as carcinoembryonic antigens (11),
`are fetal antigens, which are either not present or only present in
`small amounts on normal adult tissues. This group also contains
`antigens of unknown origin that are only defined by their reactivity
`with a monoclonal antibody (12-14).
`For targeted drug delivery to be successful, it is necessary that the
`cytotoxic agent be extremely active. Bacterial and plant toxins,
`which are some of the most cytotoxic substances known, act by
`irreversibly arresting the synthesis of protein in eukaryotic cells.
`Pseudomonas exotoxin (PE) and diphtheria toxin (DT) do this by
`enzymatically inactivating elongation factor 2, an essential compo-
`nent of protein synthesis. Ricin and other plant toxins cleave a
`glycosidic bond in 28S ribosomal RNA (rRNA), thereby destroy-
`ing the ability of ribosomes to synthesize protein. Because these
`toxins are catalysts with high turnover numbers, few molecules need
`to reach the cytoplasm to kill the target cell. These types of toxins
`enter cells by endocytosis and are processed to an active fragment
`that translocates across the cell membrane into the cytosol where the
`components ofprotein synthesis are located (Fig. 1). Ifone bypasses
`this pathway and directly injects an activated form of DT or PE into
`the cytosol, only the injection of a few molecules is necessary to kill
`the cell (15).
`Initially, agents that delivered toxins to cancer cells were con-
`structed by chemically coupling antibodies to toxins (16-19). Im-
`munotoxins have been made by the coupling of antibodies to ricin
`A chain or to several other plant toxins and to modified forms ofDT
`and PE. More recently, genetic engineering has been used to make
`recombinant toxins by fusing modified toxin genes to DNA ele-
`ments encoding growth factors or the combining regions of anti-
`bodies (20-23). These chimeric genes are then expressed in Esche-
`richia colil, from which the recombinant toxins are prepared.
`Toxins in their native form require a minimum of three biochem-
`ical functions to kill cells: cell binding, cytotoxicity, and the ability
`to translocate the toxic activity into the cytosol. An advance in
`understanding the structural basis ofeach ofthese functional regions
`occurred with the crystallization of PE and the elucidation of its
`three-dimensional structure (24). PE is a single polypeptide chain
`with a size of 66 kD that is arranged into three major structural
`domains. This arrangement suggested that each domain could be
`responsible for one function. To test this hypothesis, Hwang et al.
`(25) isolated the PE gene from Pseudomonas aeruginosa and used a
`bacterial expression system to produce the whole toxin or fragments
`of the toxin corresponding to each structural domain (25). The
`results showed that domain Ia (amino acids 1 to 252) was the cell
`binding domain, domain II (amino acids 253 to 364) was required
`
`BI
`CnOd
`pit
`
`Endocytic (7ah
`
`Fig. 1. Journey to the
`cytosol. Domain I of PE
`binds to a surface recep-
`tor and is taken into the
`endocytic compartment
`via coated pits. The low
`pH ofthe endosome and
`factors
`other
`possibly
`cause the toxin to un-
`fold. Toxin is then pro-
`in domain II. Cleavage is
`followed by reduction and the release of a 37-kD COOH-terminal fragment
`that translocates to the cytosol, and there ADP ribosylates elongation factor
`2.
`
`o
`
`-bkb
`
`F2[
`
`)
`
`(_
`i
`
`a
`
`for translocation, and domain III (amino acids 400 to 613) was
`required for the adenosine diphosphate (ADP) ribosylation and
`inactivation of elongation factor 2. A subdomain termed Ib, which
`is composed of amino acids 365 to 399, has no known function, and
`most of it can be deleted without loss of activity (26).
`The steps involved in the killing of cells by PE are shown in Fig.
`1. DT probably follows a similar pathway. Soon after toxin binding,
`the toxin-receptor complex is internalized by the pathway of
`receptor-mediated endocytosis (27, 28). The toxin travels via clath-
`rin-coated pits into endocytic vesicles where the toxin is cleaved into
`two pieces by a combination of a proteolytic step and reduction of
`a disulfide bond (29). The 37-kD fragment derived from the COOH
`portion, which contains part of domain II and all of domain III is
`translocated to the cytosol where it arrests protein synthesis and
`1 illustrates translocation
`causes cell death (29). Although Fig.
`occurring from an endocytic vesicle, recent data suggest transloca-
`tion may occur in the endoplasmic reticulum.
`Diphtheria toxin is also a single chain toxin in which the
`functional domains are arranged in the opposite order from PE,
`with the ADP ribosylating function at the NH2-terminus and the
`binding domain at the COOH-terminus (30, 31). DT made in E.
`coli is thought to be proteolytically cleaved after binding to target
`cells (32). Ricin is composed of two subunits linked together by a
`disulfide bond. The A chain contains the enzymatic activity; the B
`chain binds to galactose residues present on many different cell
`surface glycoproteins and glycolipids. Pseudomonas exotoxin, diph-
`theria toxin, and ricin are each synthesized as a single polypeptide
`chain. The chain is later proteolytically cleaved into two fragments,
`and the fragment containing the toxic enzymatic activity is translo-
`
`S-S
`
`U&
`
`Ricin
`
`S-S
`
`PE
`
`4s
`
`DT
`
`Fig.
`Strategies
`for
`2.
`making immunotoxins.
`Ricin A chain is chemi-
`cally linked to a mono-
`clonal antibody or the B
`chain of whole ricin is
`blocked with a glycopep-
`tide and modified ricin
`linked to an antibody.
`Whole PE is conjugated
`to an antibody via do-
`main I or domain I is
`deleted and then PE40 is
`conjugated to the anti-
`body. A mutant form of
`DT with reduced bind-
`ing activity is attached to
`an antibody molecule.
`
`1174
`
`SCIENCE, VOL. 254
`
`IMMUNOGEN 2146, pg. 2
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`Table 1. Clinical trials with immunotoxins. About 500 individuals have
`been treated in clinical trials. All but one of these trials used toxins
`chemically linked to antibodies (immunotoxins). Administration of the
`recombinant toxin DAB 486-IL-2 has begun in individuals with ATL and
`other IL-2 receptor-positive malignancies. The abbreviations used are as
`follows: RTA, ricin A chain; dgRTA, deglycosylated ricin A chain; CLL,
`chronic lymphocytic leukemia; T-ALL, T-cell acute lymphocytic leukemia.
`Disease
`Melanoma
`Colorectal cancer
`Breast cancer
`B cell malignancies
`CLL/T-ALL
`B cell malignancies
`B cell malignancies
`ATL
`Ovarian cancer
`ATL and other
`lymphoid tumors
`*Results of clinical trials are unpublished.
`
`Immunotoxin
`
`Xomazyme-Mel-RTA
`791T/36-RTA
`260 F9-RTA
`Anti-B4-B-ricin
`T-101-RTA
`Anti-CD22 RTA dgRTA
`Anti-CD22 Fab dgRTA
`Anti-Tac(Fv)-PE40
`OVB3-PE
`DAB 486-IL-2
`
`Phase
`1/2
`1
`1
`1/2
`1
`1
`1
`1
`1
`1
`
`References
`
`(83)
`(56)
`(54, 55)
`(37*, 84*)
`(18, 85)
`(53)
`(53)
`(86*)
`(57)
`(21*)
`
`cated into the cytosol (29, 32). PE is not cleaved (processed) until it
`enters the target cell where both proteolytic cleavage and disulfide
`bond reduction occur. In contrast, diphtheria toxin is clipped
`shortly after secretion, and ricin is cleaved within the seed where it
`is synthesized (33). In both examples the two fragments remain held
`together by a disulfide bond which is ultimately broken within the
`target cell (34).
`
`Immunotoxins
`To construct active immunotoxins, the toxin must be modified so
`that its interactions with cellular receptors are diminished or abol-
`ished (Fig. 2). As a consequence, toxin entry is mediated by
`antibody binding. With ricin, this is accomplished through removal
`of the B chain (35), blockage of the galactose binding site (36) or
`attachment of a galactose-rich carbohydrate to the B chain (37).
`With PE, this modification occurs when the antibody is coupled to
`domain I ofPE, which interferes with the binding ofdomain I to the
`PE receptor (38), or when domain I (amino acids 1 to 252) is
`genetically deleted and coupled to antibody to domain
`(39). With
`DT, this modification occurs through the mutation of a key amino
`acid in the binding domain near the COOH-terminus or through
`the removal of a portion of the COOH-terminus that is responsible
`for binding to cells (40, 41).
`Ricin is purified from castor beans and has been available for
`immunotoxin production for many years. More recently, a recom-
`binant form of the A chain has been produced in E. coli (42). PE has
`been obtained from the culture medium of P. aeruginosa, but both
`the whole toxin and mutant forms have been produced in E. coli (20)
`as have been mutant forms of DT.
`The activity of an immunotoxin is initially assessed by measuring
`its ability to kill cells with target antigens on their surfaces. Because
`toxins act within the cell, receptors and other surface proteins that
`naturally enter cells by endocytosis usually make good targets for
`immunotoxins, but surface proteins that are fixed on the cell surface
`do not. However, ifseveral antibodies recognizing different epitopes
`on the same cell surface protein are available, it is useful to test them
`all, because some, perhaps by producing a conformational change in
`the target protein's structure, may induce its internalization ordirect
`its intracellular routing to an appropriate location for toxin translo-
`cation (43, 44). Also, it is possible to induce internalization of a
`target surface protein if the immunotoxin contains a form of PE or
`ricin in which the binding of the toxin moiety to its receptor,
`
`22 NOVEMBER 1991
`
`although weakened by chemical modification, still occurs and pro-
`motes internalization since toxin receptors are efficiently internalized
`(14, 37, 41).
`Many immunotoxins produce selective killing in cell culture, but
`only a few of these have been able to cause substantial or complete
`tumor regression in animals. Ricin A chain coupled to antibodies
`recognizing B cell specific antigens have caused complete regression
`of B cell lymphomas in mice (45). When antibodies to carcinomas
`are used, only partial responses have been observed (46-48). Re-
`gression of human carcinomas growing in immunodeficient mice
`has been achieved by treatment with monoclonal antibodies reacting
`with ovarian, colon, and breast cancers coupled either to PE itselfor
`to PE40, a mutant form ofPE in which the cell binding domain was
`deleted (14, 49, 50). One of these, B3-PE40, causes complete
`regression of human tumors growing in mice (51).
`Several immunotoxins have been developed and approved for
`human trials. Two different kinds of trials have been conducted.
`The first involves the ex vivo addition of immunotoxins to
`harvested bone marrow to eliminate contaminating tumor cells
`before reinfusion in patients undergoing autologous bone marrow
`transplantation. A variety of antibodies, linked to ricin or ricin A
`chain, including anti-CD5 and anti-CD7, have been used for this
`purpose (52). The second kind of trial involves the parenteral
`administration of immunotoxins, either regionally (such as the
`peritoneal cavity) or systemically to patients with cancer. These
`have been primarily Phase 1 and 2 trials in patients in which
`conventional treatments have failed, and the patients have a large
`tumor burden. A list of clinical trials, either completed or still in
`progress, is provided in Table 1. Definite responses have been
`noted with lymphomas, and these trials will be expanded (53). So
`far, the antibodies used for the preparation of immunotoxins to
`treat carcinomas or other solid tumors have been found to react
`with important normal human tissues (such as neural tissue and
`bone marrow) and produce dose-limiting toxicity without signif-
`icant clinical responses (54-57).
`
`Geneticaily Engineered Recombinant Toxins
`The production ofimmunotoxins by chemical coupling methods is
`expensive because it requires large amounts of antibody and toxin.
`Furthermore, the chemical conjugation methods used produce heter-
`ogeneous products, and antigen binding is often affected by chemical
`derivatization. It has been possible to overcome these difficulties and
`to create cytotoxic agents by genetic engineering. Both PE and DT
`have been used to make recombinant toxins in E. coli (20, 23).
`Ricin-based molecules have been difficult to produce probably because
`the A chain ofthe plant toxin must be attached to the cell recognition
`domain by a disulfide bond, and disulfide-linked subunits are difficult
`to produce in bacteria. The addition ofa proteolytic cleavage sequence
`may help to overcome this difficulty (58).
`The x-ray crystallographic structure of PE has been used as a
`guide for the synthesis of genetically engineered recombinant
`toxins (Fig. 3). The specific binding of PE to target cells occurs
`through an interaction of domain I with cellular PE receptors (59,
`60). The junction of domain I with domain II occurs between
`Glu2`2 and Gly253. Therefore, in constructing recombinant toxins,
`domain I was deleted, and the COOH-terminal-amino acid of
`various growth factors and other targeting molecules were fused
`directly to Gly253 of PE (occasionally a few additional amino acids
`have been added as a link between the COOH-terminus of the
`growth factor and Gly253 of PE to make cloning more conve-
`nient). One widely studied molecule is TGF-a-PE40, which was
`constructed by replacing domain I of PE with transforming
`
`ARTICLES
`
`1175
`
`IMMUNOGEN 2146, pg. 3
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`Fig. 3. Strategies for making recombinant
`toxins. The DNA coding for domain I of
`PE is removed and replaced with cDNAs
`for ligands such as TGF-oa, IL-2, IL-6, and
`single chain antibodies. The cDNAs are
`placed at the 5' end of each construction
`and preserve the relative position of the
`binding function to the other functional
`domains of PE. Binding by DT is mediat-
`ed by sequences at the 3' of its structural
`gene. Therefore, ligands are added to a
`truncated form of DT at the 3' end.
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`
`7 _i
`
`a
`
`growth factor alpha (TGF-ot). In this chimeric toxin, the 23-U)
`domain I is replaced by the 6-kD growth factor (Fig. 3), to
`produce a chimeric toxin that selectively binds to and kills cells
`with epidermal growth factor (EGF) receptors. The structure of
`the plasmid encoding TGF-c-PE40 is shown in Fig. 4 with the
`cDNA encoding TGF-a inserted adjacent to the PE40 gene. The
`expression vector used for the production of TGF-a-PE40 and
`other PE-based chimeric toxins in E. coli contains the bacterio-
`phage T7 promoter, an efficient ribosome binding site and an Nde
`I site (CATATG), which encodes a methionine initiation codon
`where targeting ligands can be conveniently inserted. The gene
`encoding the phage T7 polymerase is inserted into the E. coli
`chromosome next to a lac promoter so that it can be induced by
`the addition of isopropylthiogalactoside (IPTG) (61).
`TGF-a-PE40, like other chimeric toxins made in E. coli,
`accumulates in large amounts within the cell in insoluble aggre-
`gates (inclusion bodies). After cell disruption, inclusion bodies are
`easily isolated and can contain up to 90% recombinant protein in
`an insoluble form. The protein is then dissolved in a strong
`denaturant such as 7 M guanidine-HCl, renatured, and can be
`
`TGF-a
`o
`Met Gly Val
`CATATGGGAGTG
`
`TGF-a
`
`PE40---- -
`(ix)
`1
`2
`3
`253
`Glu Glu Gly
`Leu Ala Met Al
`ATGGCCGAAGAGGGC
`
`E4
`
`1
`
`I
`
`1
`
`7c F
`
`pVC 387<
`I~~~~~~~~Ampr
`
`I-l
`'
`Fig. 4. Expression of recombinant toxins from plasmids containing the T7
`promoter. Production of PE recombinant toxins is driven by T7 polymerase,
`which is made by induction of the lactose operon situated on the chromo-
`some of E. coli BL21 (XDE3). The polymerase acts on the T7 promoter
`immediately upstream of the structural gene for the recombinant toxin.
`TGF-a is shown fused with PE40. This construction is inserted into an Nde
`I site, which provides the first methionine codon to begin the translation of
`the fusion protein. Additional amino acids are present at the junction of two
`structural genes (79).
`
`1176
`
`Table 2. Targets for recombinant toxins.
`
`Surface target
`EGF receptor
`
`IL-2 receptor
`
`Relevant cancers
`
`Lung, head and neck,
`bladder,
`glioblastoma,
`breast
`ATL
`T cell lymphomas
`
`IL-6 receptor
`
`Myeloma, hepatoma
`
`Erb-B2 protein
`Cancer-associated
`antigens
`
`Breast, ovarian, bladder
`Carcinomas
`
`Recombinant toxins
`TGF-a-PE40 (88)
`
`IL-2-PE40 (64)
`DT-IL-2 (23)
`Anti-Tac(Fv)-PE40 (22)
`DT-anti-Tac(Fv) (87)
`IL-6-PE40 C 2)
`IL-6-PE664 lu (89)
`Anti-ErbB2(Fv)-PE40 (90)
`B3(Fv)-PE40 (80)
`
`purified to near homogeneity in two or three steps by conventional
`column chromatographic methods (62). TGF-a-PE40 binds to
`EGF receptor-containing cells with about the same affinity as
`TGF-ct, and its toxicity on these cells is directly related to the
`number of receptors present.
`
`Recombinant Toxins and Their Targets
`The EGF receptor (EGF-R) has been the subject of intense
`study and shown to act as an oncoprotein when it is overexpressed
`in normal cells (2). TGF-t-PE40 is extremely cytotoxic to cells
`that contain EGF receptors and has been shown to have an
`antitumor effect in animals bearing tumors that have more EGF
`receptors than normal cells (50, 63). Clinical trials have just begun
`in which TGF-a-PE40 is instilled into the urinary bladder to treat
`superficial bladder cancer. Recombinant toxins have now been
`made by combining IL-2, IL-4, IL-6, IGF-1, and acidic fibroblast
`growth factors (FGF) with PE40 (62, 64-67) and DT with IL-2
`and melanocyte-stimulating hormone (21, 68) (Fig. 3). Each was
`cytotoxic to cell lines containing the appropriate receptors. Fur-
`thermore, anti-tumor activity in animals bearing human cancers
`has been demonstrated with TGF-a-PE40, IL-2-PE40, and
`IL-6-PE40 (69, 70). Another way to assess efficacy of such agents
`is to test them against cancer cells directly isolated from patients.
`DT-IL-2 has activity against fresh leukemic cells from patients
`with ATL (71) and is now being evaluated in Phase 1 trials in
`individuals with various lymphoid malignancies. Table 2 summa-
`rizes some tumor types that contain large numbers of receptors
`and may serve as targets for recombinant toxin therapy.
`Although growth factors fused to toxins have proved to be
`effective cytotoxic agents, a concern has been raised that some of
`these might promote tumor growth in certain circumstances, for
`example, if less than a full toxic dose were administered (65, 72,
`73). Because antibodies to growth factor receptors do not usually
`have agonist effects, the antigen combining region of these
`antibodies can be used for targeting. This approach takes advan-
`tage of the finding that the variable regions of the light and heavy
`chains of antibodies can be combined into a single chain form that
`retains high affinity binding to antigen (74-76). Accordingly, a
`complementary DNA (cDNA) from the antibody to Tac (anti-
`Tac) was used to construct an Fv fragment that was fused to PE40
`(22) (Fig. 3). The Fv portion of an antibody molecule is composed
`of two variable regions of the light and heavy chains and these can
`be held together by a linking peptide to make a single chain Fv.
`This single-chain recombinant toxin anti-Tac(Fv)-PE40, which is
`under preclinical development has potent cell-killing activity
`against cells with IL-2 receptors and against cells directly isolated
`SCIENCE, VOL. 254
`
`IMMUNOGEN 2146, pg. 4
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`from ATL patients (77). The approach, of combining single chain
`antibodies with PE40, has been extended to several other anti-
`bodies, including an antibody to the human transferrin receptor
`(78), OVB3 (79), and B3 (80). The antibody to B3, which binds
`to a carbohydrate antigen expressed on the surface of many
`carcinomas (81), has been used to make a single-chain recombi-
`nant toxin that causes the complete regression of human tumors in
`mice (80).
`
`Problems and Prospects
`Toxins are foreign proteins and highly immunogenic. Therefore,
`in the absence of immunosuppression, neutralizing antibodies de-
`velop about 10 days after exposure to toxin (57, 82) and antibodies
`to DT already exist in most individuals who have received immuni-
`zations with diphtheria, pertussis, and tetanus (DPT). Animal
`studies have shown that immunotoxins and recombinant toxins act
`quickly so that tumors regress in a few days. Similar rapid responses
`should occur in humans. Nevertheless for long-term therapy, con-
`comitant administration of immunosuppressive agents will be nec-
`essary to prolong the treatment period or the same antibody or
`ligand can be used to target different toxins. Many of the patients
`being treated in the current clinical trials have received extensive
`chemotherapy and are severely immunosuppressed so that the
`treatment period can be extended. In this review, we have discussed
`the utility of three toxins that are under clinical development, but
`there are still concerns associated with their use. Many people in the
`Western world have neutralizing antibodies to DT as a result of
`childhood immunization, and this has caused concern about the
`utility of recombinant toxins containing DT. Ricin A chain and
`blocked ricin both make active immunotoxins, but the recombinant
`molecules produced so far have low cytotoxic activity. The advan-
`tage of PE is that its structure is known. And, by means of this as a
`guide, the cell binding domain has been successfully replaced with
`more than a dozen different ligands including several different single
`chain antibodies. Furthermore, less than 3 percent of humans have
`preexisting antibodies to PE. In the next several years, we anticipate
`that the major role of immunotoxins and recombinant toxins will be
`in the adjuvant setting for the treatment of metastatic disease that
`remains after surgery or radiation therapy.
`
`REFERENCES AND NOTES
`1. D. FitzGerald and I. Pastan, J. Nati. Cancer Inst. 81, 1455 (1989).
`2. T. J. Velu et al., Science 238, 1408 (1987).
`3. M. Kawano et al., Nature 332, 83 (1988).
`4. K. E. Hellstrom and I. Hellstrom, FASEBJ. 3, 1715 (1989).
`5. F. J. Hendler and B. W. Ozanne, J. Clin. Invest. 74, 647 (1984).
`6. N. R. Jones, M. L. Rossi, M. Gregoriou, J. T. Hughes, Cancer 66, 152 (1990).
`7. J. L. T. Lau, J. E. J. Fowler, L. Ghosh, J. Urol. 139, 170 (1988).
`8. W. A. Dunn, T. P. Connolly, A. L. Hubbard, J. Cell Biol. 102, 24 (1986).
`9. T. A. Waldmann, Cell Immunol. 99, 53 (1986).
`10. M.-A. Ghetie et al., Cancer Res. 48, 2610 (1988).
`11. R. Muraro et al., ibid. 45, 5769 (1985).
`12. A. E. Frankel, D. B. Ring, F. Tringale, M. S. Hsieh, J. Biol. Response Mod. 4, 273
`(1985).
`13. N. Varki, R. A. Reisfeld, L. E. Walker, Cancer Res. 44, 681 (1984).
`14. M. C. Willingham, D. J. FitzGerald, I. Pastan, Proc. Nad. Acad. Sci. U.SA. 84,
`2474 (1987).
`15. M. Yamaizumi, E. Makada, T. Uchida, Y. Okada, Cell 15, 245 (1978); I. Pastan
`and D. FitzGerald, unpublished data.
`16. E. S. Vitetta, R. J. Fulton, R. D. May, M. Till, J. W. Uhr, Science 238, 1098
`(1987).
`17. I. Pastan, M. C. Willingham, D. J. FitzGerald, Cell 47, 641 (1986).
`18. A. A. Herder et al., J. Biol. Response Mod. 7, 97 (1988).
`19. D. J. Neville, Crit. Rev. Therap. Drug. Carrier. Syst. 2, 329 (1986).
`20. I. Pastan and D. FitzGerald, J. Biol. Chem. 264, 15157 (1989).
`21. D. P. Williams et al., Protein Eng. 1, 493 (1987).
`
`22. V. K. Chaudhary et al., Nature 339, 394 (1989).
`23. J. R. Murphy, Cancer Treat. Res. 37, 123 (1988).
`24. V. S. Allured, R. J. Collier, S. F. Carroll, D. B. McKay, Proc. Natl. Acad. Sci.
`U.SA. 83, 1320 (1986).
`25. J. Hwang, D. J. FitzGerald, S. Adhya, I. Pastan, Cell 48, 129 (1987).
`26. C. B. Siegall, V. K. Chaudhary, D. J. FitzGerald, I. Pastan, J. Biol. Chem. 264,
`14256 (1989).
`27. D. J. P. FitzGerald, R. E. Morris, C. B. Saelinger, Cell 21, 867 (1980).
`28. R. E. Morris and C. B. Saelinger, Surv. Synth. Pathol. Res. 4, 34 (1985).
`29. M. Ogata, V. K. Chaudhary, I. Pastan, D. J. FitzGerald, J. Biol. Chem. 265,
`20678 (1990).
`30. D. M. J. Neville, T. H. Hudson, Annu. Rev. Biochem. 55, 195 (1986).
`31. A. M. J. Pappenheimer, ibid. 46, 69 (1977).
`32. D. P. Williams et al., J. Biol. Chem. 265, 20673 (1990).
`33. S. M. Harley and J. M. Lord, Plant Sci. 41, 111 (1985).
`34. S. Olsnes and K. Sandvig, Cancer Treat. Res. 37, 39 (1988).
`35. H. E. Blythman et al., Nature 290, 145 (1981).
`36. P. E. Thorpe et al., Eur. J. Biochem. 140, 63 (1984).
`37. J. M. Lambert et al., Biochemistry 30, 3234 (1991).
`38. D. J. FitzGerald, M. C. Willingham, I. Pastan, Cancer Treat. Res. 37, 161 (1988).
`39. T. Kondo, D. FitzGerald, V. K. Chaudhary, S. Adhya, I. Pastan, J. Biol. Chem.
`263, 9470 (1988).
`40. L. Greenfield, V. G. Johnson, R. J. Youle, Science 238, 536 (1987).
`41. M. Colombatti, L. Greenfield, R. J. Youle, J. Biol. Chem. 261, 3030 (1986).
`42. M. Piatak et al., ibid. 263, 4837 (1988).
`43. R. D. May, H. T. Wheeler, F. D. Finkelman, J. W. Uhr, E. S. Vitetta, Cell
`Immunol. 135, 490 (1991).
`44. 0. W. Press, P. J. Martin, P. E. Thorpe, E. S. Vitetta, J. Immunol. 141, 4410
`(1988).
`45. K. A. Krolick, J. W. Uhr, S. Slavin, E. S. Vitetta, J. Exp. Med. 155, 1797 (1982).
`46. T. W. Griffin et al., Cancer Res. 47, 4266 (1987).
`47. D. J. FitzGerald et al., ibid., p. 1407.
`48. H. Masui, H. Kamrath, G. Apell, L. L. Houston, J. Mendelsohn, Cancer Res. 49,
`3482 (1989).
`49. J. K. Batra et al., Proc. Natl. Acad. Sci. U.S.A. 86, 8545 (1989).
`50. L. H. Pai, M. G. Gallo, D. J. FitzGerald, I. Pastan, Cance