- Email: [email protected]
Targeting diphtheria toxin to growth factor receptors
seminars in
CANCER BIOLOGY, Vol 6, 1995: pp 259–267
Targeting diphtheria toxin to growth factor receptors John R. Murphy and Johanna C. vanderSpek
negative host. The second experiment addressed the possibility of creating ‘new’ toxins that would be selectively active and eliminate only those cells expressing target receptor on their surface. Although these experiments were approved by the RAC, this research was put on administrative hold for almost four years, in part, because of its controversial nature. By the fall of 1985, we were finally allowed to proceed with only those experiments involving the genetic substitution of the native diphtheria toxin receptor binding domain with α-MSH at the National Institutes of Allergy and Infectious Diseases (NIAID) BL-4 facility at Frederick, Maryland.
Biochemical, genetic and X-ray crystallographic analysis of diphtheria toxin have demonstrated that the native toxin is composed of three structural domains that function in an ordered fashion to intoxicate a eukaryotic cell. With the knowledge that, if delivered to the cytosol, a single molecule of the catalytic domain is lethal for the cell, we have used recombinant DNA methods to genetically replace the native toxin receptor binding domain with a series of growth factors. The resulting diphtheria toxin-related cytokine fusion proteins, or fusion toxins bind to their respective receptors, are internalized by receptor-mediated endocytosis, and efficiently eliminate target cell populations by the adenosine diphosphate ribosylation of elongation factor 2. Based upon the results of preclinical studies, DAB486IL-2, DAB389IL-2 and DAB389EGF have, or are in the process of being evaluated in Phase I/II clinical trials. To date, administration of the diphtheria toxin-based fusion proteins targeted toward the high affinity IL-2 receptor have been found to be safe, well tolerated, and capable of inducing remission in refractory hematologic malignancies.
The diphtheria toxin platform for genetic engineering The choice of diphtheria toxin as the toxophore for receptor binding domain substitution was based upon a number of earlier observations. In their classic study, Uchida, Gill and Pappenheimer1 demonstrated that the structural gene for diphtheria toxin was carried by corynebacteriophage β. This study also provided the foundation for subsequent studies on the structure function relationships of diphtheria toxin by demonstrating that the enzymatically active A fragment was positioned on the N-terminal end of the toxin; whereas, the receptor binding domain of the toxin was carried on fragment B. Shortly thereafter, Murphy et al 2 used β-phage DNA to program S-30 extracts of E. coli and demonstrated, in this coupled transcription translation system, that biologically active diphtheria toxin could be synthesized in vitro. Using [3H]-Triton X-100, Boquet et al 3 clearly demonstrated that fragment B of both diphtheria toxin and the non-toxic mutant CRM45 carried a hydrophobic domain which had properties of integral membrane proteins. These investigators postulated that the role of this hydrophobic domain was to facilitate the delivery of fragment A across the eukaryotic cell membrane and into the cytosol. Thus, it was known quite early that native diphtheria toxin was a three domain protein consisting of the enzymatically active domain (fragment A), the hydrophobic
Key words: cytotoxicity / diphtheria toxin / growth factor / psoriasis / receptor ©1995 Academic Press Ltd
ALMOST 15 years ago we submitted a proposal to the Recombinant DNA Advisory Committee (RAC) to conduct two experiments under Biosafety Level 4 (BL4). The first experiment was to clone the structural gene for diphtheria toxin in Escherichia coli K-12 and, based upon the assumption that biologically active diphtheria toxin would be expressed and secreted into the periplasmic space, the second experiment proposed the genetic substitution of the native diphtheria toxin receptor binding domain with the polypeptide hormone α-melanocyte stimulating hormone (α-MSH). The initial experiment was designed to test the feasibility of expressing a potent toxin from a Gram positive organism in a heterologous Gram From Evans Department of Clinical Research and Department of Medicine, Boston University Medical Center Hospital, Boston, MA 02118, USA ©1995 Academic Press Ltd 1044-579X/95/050259 + 09 $12.00/0
259
J. R. Murphy and J. C. vanderSpek subsequent to receptor binding by cell surface associated proteases. In either case, it is clear that processing must occur before the catalytic domain is translocated to the cytosol. The protease recognition sequence of diphtheria toxin is Arg190-Val191Arg192-Arg193 which has been shown to be a site for the endoprotease furin.11,12 Additional support for a furin-mediated cleavage of the toxin as an essential step in the intoxication process comes from a sitedirected mutational analysis of DAB486IL-2, in which destruction of the furin recognition site resulted in a marked loss of cytotoxic activity in the fusion toxin.13 More recently, it has been shown that furin can be used to specifically nick intact diphtheria toxin in vitro.14 Tsuneoka et al 15 have shown that LoVo cells, which do not produce furin are not sensitive to the action of diphtheria toxin; however, upon transfection of LoVo cells with the gene encoding furin, the cells become sensitive to the toxin. Since it has been shown that furin cycles from the trans-Golgi to the membrane,16 it is almost certain that intact diphtheria toxin, and the diphtheria-related fusion toxins are processed on the cell surface by furin.
domain (N-terminal portion of fragment B), and the receptor binding domain (C-terminal portion of fragment B). The process by which diphtheria toxin intoxicates sensitive eukaryotic cells involves at least the following steps: (i) binding of the toxin to its cell surface receptor, (ii) activation of the catalytic domain by a proteolytic cleavage (‘nicking’) of the toxin in a sensitive exposed 14 amino acid loop that is subtended by Cys186 and Cys201, (iii) internalization of the bound toxin into endosomes by receptor-mediated endocytosis, and following acidification of the endocytic vesicle, (iv) the facilitated delivery of the catalytic domain across the endocytic vesicle membrane and into the cytosol. Once delivered to the cytosol fragment A rapidly catalyses the adenosine diphosphate ribosylation of elongation factor 2 which results in the inhibition of protein synthesis and subsequent death of the cell.
Receptor binding domain The first step in the intoxication process is the specific binding of diphtheria toxin to its cell surface receptor. Middlebrook et al 4 were able to correlate the apparent sensitivity of a given cell line to diphtheria toxin with the number of receptors on the cell surface. The initial localization of the diphtheria toxin receptor binding domain to the carboxy-terminal region of fragment B was based upon the findings that CRM45, a pre-mature chain termination mutant of the toxin which lacked the C-terminal 15,000 dalton region, failed to block the toxic activity of diphtheria toxin on cells.1,5 In addition, observations made by many investigators using different approaches strongly suggested that the functional native receptor binding domain was positioned in the carboxy-terminal 50 amino acids of the toxin.6-10
Transmembrane domain It has been known for over a decade that once diphtheria toxin is bound to its cell surface receptor, it is internalized into the cell by receptor-mediated endocytosis.17,18 Early endocytic vesicles are known to be acidified by specific vesicular ATPases to an average pH value of 6.2.19,20 It is well known that diphtheria toxin must pass through an acidic compartment in order to deliver the catalytic domain to the cytosol. Over a decade ago, it was recognized that under acidic conditions diphtheria toxin and CRM45 will spontaneously insert into the plane of lipid bilayers and form channels.21,22 Moreover, the diameter of the channel has been reported to be 18 A˚ 21,23 which is large enough for a denatured, extended fragment A to pass through and be delivered to the cytosol. Importantly, diphtheria toxin-induced channels have been observed in both Vero and CHO cell membranes following a low pH-pulse.24,25 The channel formation that has been observed in both artificial and cellular membranes involves the insertion of transmembrane domain α-helices. O’Keefe et al 26 found that the introduction of the E349K mutation, which is positioned in the short loop connecting the transmembrane helices 8 and 9, resulted in both a decrease in the cytotoxic potency and channel-forming activity of the mutant toxin.
Processing/activation domain Once bound to its cell surface receptor, intact diphtheria toxin must be processed into an active two chain protein through proteolytic processing, or nicking. Intact diphtheria toxin has an exposed serine-protease sensitive 14-amino acid loop that is subtended by a disulfide bond between Cys185 and Cys201. As shown in Figure 1, the cleavage domain of the toxin is positioned between the catalytic and transmembrane domains. Nicking, or processing of the toxin may occur either prior to the toxin binding to its cell surface receptor by serum proteases or 260
Targeting diphtheria toxin These investigators proposed that following protonation of E349, and possibly D325 in the endocytic vesicle, transmembrane helices 8 and 9 would spontaneously insert into the plane of the membrane. Further, upon exposure of the connecting loop to the
neutral pH environment of the cytosol, E349 would become deprotonated and its charge would lock, or anchor this α-helical hairpin in the membrane. In an analogous fashion, Falnes et al 27 reported that the diphtheria toxin-related site-directed mutant D318K
Figure 1. Ribbon diagram of the X-ray crystal structure of native diphtheria toxin34 as modified by Bennett et al.52 The catalytic domain (red), transmembrane domain (magenta), and receptor binding domain (blue) are shown. N, N-terminal end of the toxin; PSL, 14 amino-acid protease sensitive loop which separates the catalytic from transmembrane and receptor binding domains; C, C-terminal end of the toxin. The ribbon diagram was generated using MOLESCRIPT.63
261
J. R. Murphy and J. C. vanderSpek complete unfolding of the catalytic domain comes from the observations of Falnes et al. 33 Using the X-ray crystal structure of diphtheria toxin,34 Falnes et al 33 showed the introduction of paired cysteine residues in the catalytic domain, which were likely to form disulfide bridges, resulted in the formation of mutants with decreased cytotoxic activity. While it is clear that the disulfide bond between the catalytic and transmembrane domains (Cys186– Cys201) must be reduced in order to release the catalytic domain, the precise cellular location of where this reduction occurs is unknown. Moskaug et al 35 provided evidence that the reduction of the disulfide bond occurs either when the transmembrane domain becomes buried in the lipid core or when the bond is exposed to the cytosol. More recently, Papini et al 36 demonstrated that the reduction of the disulfide bond occurs after the low-pH insertion of the transmembrane domain into an early endosomal compartment. Since membrane-impermeant sulfhydryl blockers inhibit the action of diphtheria toxin, it has been proposed that the reduction of the disulfide bond between the catalytic and transmembrane domains could happen either at the level of the cell surface or in the lumen of the endocytic vesicle. In support of this hypothesis, Mandel et al 37 have provided evidence that protein disulfide isomerase (PDI) plays a major role in the diphtheria toxin mediated intoxication of Vero cells. Since PDI has been reported to be primarily localized in the endoplasmic reticulum38 and direct evidence demonstrating that PDI plays a role in the intoxication process is lacking, this hypothesis awaits further experimental support.
[D318 is positioned in the loop connecting transmembrane α-helices 5 and 6] also had reduced cytotoxic potency and impaired ability to form channels in artificial membrane bilayers. vanderSpek et al 28 have shown that an intact transmembrane helix 9 is also required for the selective toxicity DAB389IL-2. Inframe deletion mutants that lack the C-terminal portion of transmembrane helix 9 had markedly reduced cytotoxic potency and were incapable of forming stable channels in artificial membranes. The amino-terminal region of the transmembrane domain is amphipathic and has been shown to be structurally homologous to apolipoprotein A1.29 The introduction of mutations into this portion of diphtheria toxin have been shown to result in mutants with a decreased ability to bind to the diphtheria toxin receptor.30 In contrast, mutations in this region of the fusion toxin molecule do not effect receptor binding affinity; however, they may result in non-toxic mutants in which neither receptor binding activity nor channel formation are impaired.31 Since these mutants bind with high affinity and are capable of forming channels in artificial membranes, they define a hitherto unknown step in the intoxication process.
Catalytic domain Relatively little is known of the molecular details involving the translocation of the catalytic domain of diphtheria toxin through the endocytic vesicle membrane and into the cytosol. Moreover, most of our current understanding of the mechanism by which the catalytic domain enters the cytosol is based upon indirect observations. For example, evidence supporting the hypothesis that the catalytic domain must completely unfold comes from the early observations that the pore formed by insertion of the transmembrane domain into the membrane is approximately 18 A˚ in diameter; only large enough to allow the passage of a completely unfolded catalytic domain. More recently Wiedlocha et al 32 described the construction of a fusion protein in which acidic fibroblast growth factor (aFGF) was linked to the N-terminus of the catalytic domain. While this fusion protein was as toxic toward Vero cells as the native toxin, the addition of heparin to the assay medium (i.e. heparin induces a tight folding of aFGF) resulted in a complete loss of toxicity. These investigators reasoned that the heparin-induced tight folding of the aFGF component of the fusion protein caused a block in the translocation of the catalytic domain into the cytosol. Perhaps the most compelling evidence in support of
Diphtheria toxin-related growth factor fusion proteins In 1980, we39 began to investigate the role that the hydrophobic domain of fragment B played in the delivery of fragment A to the cytosol of target cells. This study was prompted by the observation that an epidermal growth factor (EGF) ricin A-chain conjugate toxin was exquisitely cytotoxic for cells bearing the EGF receptor; whereas, the analogous conjugate toxin assembled with fragment A of diphtheria toxin was essentially devoid of activity.40 We reasoned that the failure of diphtheria toxin fragment A-based conjugate to be active was likely due to either the lack of a hydrophobic domain that facilitated the entry of the catalytic domain to the cytosol or to stearic 262
Targeting diphtheria toxin of the fusion toxins might be useful in the treatment of human disease. Murphy et al 7 described the construction and properties of a fusion protein, DAB486α-MSH, in which the native diphtheria toxin receptor binding domain was replaced with α-melanocyte stimulating hormone (α-MSH). This construct employed a unique SphI restriction endonuclease site in the diphtheria toxin structural gene such that amino acid 486 served as the fusion junction between diphtheria toxin-related sequences and α-MSH. Despite the fact that DAB486α-MSH was subject to marked proteolytic degradation in E. coli, we were able to partially purify sufficient levels of the fusion toxin to demonstrate α-MSH receptor specific toxicity. Importantly, we were able to demonstrate by guinea pig and mouse challenge experiments conducted under BL-4 containment, that the recombinant E. coli expressing DAB486α-MSH were avirulent. This data supported the argument that further construction of novel recombinant diphtheria toxin-related cytokine fusion genes could safely be conducted in our laboratory in Boston under BL-3 containment. Since DAB486α-MSH was subject to marked proteolytic degradation, and the protease(s) sensitive sites appeared to be close to the fusion junction between diphtheria toxin and α-MSH, we reasoned that a fusion toxin constructed with a growth factor of larger mass might provide stearic hinderance and thereby minimize degradation. In fact this was found to be the case, and much of our current understanding of the diphtheria toxin-related fusion toxins comes from DAB486IL-2 and DAB389IL-2 in which the native receptor binding domain of the toxin was replaced with interleukin 2.41-44 As shown in Table 1, a variety of growth factors have been used to replace the native receptor binding domain of the toxin. In all instances, these fusion toxins have been shown to be selectively
constraints imposed by the EGF portion of the conjugate. Accordingly, we modified the imidazole ring of the central histidine of thyrotropin releasing hormone (TRH; pyro-Glu-His-Pro-NH2) by the addition of acetylcystamine. Once modified, acetylcystaminyl-THR was then coupled through a disulfide bond to either CRM26 (fragment A) or CRM45. While both CRM26-TRH and CRM45-TRH were found to specifically bind to the TRH receptor on CH3 rat pituitary cells, only the CRM45-TRH conjugate was cytotoxic (IC50 = 3 3 10–9M). In marked contrast, CRM26-TRH was devoid of activity at concentrations greater than 10–7 M. These experiments demonstrated that the hydrophobic domain of fragment B was essential in order to facilitate the delivery of fragment A to the cytosol. The most troubling aspect encountered in the assembly of the CRM45-TRH conjugates was that the specific toxicity (IC50/mg conjugate) varied widely from preparation to preparation. This variation was presumably due to differences in the extent of acetylcystaminylation of native TRH, differences in either the extent of modification of CRM45 with N-succinimidyl-3-(2-pyridyldithio)-proprionate (SPDP), and/or in the extent of coupling of the modified proteins into the hormone toxin conjugate. In order to overcome the uncertainties associated with chemical conjugation, we turned to the methodologies of protein engineering and recombinant DNA to assemble fusion genes in which the native receptor binding domain of diphtheria toxin was replaced with specific ligands. From the outset, the prospect of using recombinant DNA methods to assemble the structural genes encoding bacterial toxin growth factor fusion proteins, or fusion toxins, offered significant advantages over chemical conjugation in the assembly of chimeric proteins. Most importantly, the fusion junction, or point at which the substitute receptor binding domain was linked to the toxin fragment, was precisely determined. Subsequent expression in recombinant E. coli then resulted in the synthesis of a single gene product rather than the mixture of isomeric forms that result from chemical conjugation of two proteins. Thus, we reasoned that by in-frame deletion analysis, it should be possible to determine the diphtheria toxin fragment B (transmembrane domain) structure required for the efficient delivery of fragment A through the membrane and into the cytosol. Furthermore, with the proviso that the distribution of the targeted receptor was limited to subsets of cells that were involved in a pathogenic process, we envisioned that at least some
Table 1. Diphtheria toxin-based cell receptor-targeted fusion proteins Fusion toxin DAB486α-MSH DAB389α-MSH DAB486IL-2 DAB389IL-2 DAB389IL-4 DAB389IL-6 DAB389IL-7 DAB389EGF DAB486CD4
263
Receptor
Cytotoxicity (IC50)
Ref
α-MSH α-MSH IL-2 IL-2 IL-4 IL-6 IL-7 EGF HIV gp120
n.d. 3×10–11M 1×10–11M 1×10–12M 2×10–10M 2×10–11M 1×10–10M 1×10–11M 1×10–9M
7 47 41 64 46 49 50 48 45
J. R. Murphy and J. C. vanderSpek Table 2. Cytotoxicity, binding affinity, and planar lipid bilayer conductance of DAB389IL-2 and in-frame deletion mutants (modified from ref 51) Fusion toxin DAB389IL-2 DAB (∆381–387)389IL-2 DAB (∆369–387)389IL-2
Cytotoxicity (IC50×10–12M)
Binding affinity (Kd×10–12M)
Conductance (pS)
3 70 6,000
6.6 8.0 9.1
42 40 *33*
*Unstable channels.
cokinetics of DAB486IL-2 in patients with refractory hematologic malignancies.55-57 These early studies employed DAB486IL-2 as the experimental therapeutic in order to establish the ‘proof of principal’ that the diphtheria toxin-based fusion toxins would show some degree of efficacy in the treatment of human disease. In fact, several of these clinical studies were initiated prior to the isolation and characterization of the more potent DAB389IL-2 form of the fusion toxin. In order to be eligible for study entry, patients had to present with a hematologic malignancy that was refractory to standard treatment regimens. Patients of either sex over the age of 17 years could be treated if they had adequate hepatic and renal function and a Karnofsky performance score of 70% or greater. The initial clinical trials were designed as a three patient cohort dose escalation in which single and multiple doses of the fusion toxin were administered by intravenous injection, either as a bolus, or 90-minute infusion. The initial dose was 700 ng/kg per day and was escalated to 400 µg/kg per day. Patients were monitored for adverse effects and response and were scored according to National Cancer Institute criteria. Patients with no evidence of disease for at least four weeks were classified as complete responders (CR); patients whose tumor burden decreased by ≥ 50% for at least four weeks were classified as partial responders (PR); and patients whose tumor burden decreased by ≥ 50% were classified as minor responders. The intravenous administration of DAB486IL-2 was found to be well tolerated at all dose levels. The adverse effects were generally mild and included nausea/vomiting, hypersensitivity, fever/malaise/ chills, and elevations in serum hepatic transminases. Renal insufficiency defined the maximum tolerated dose at levels of DAB486IL-2 above 400 µg/kg per day. Importantly, in all instances the adverse effects were transient, not cumulative, and did not preclude repeated administration of the fusion toxin to patients who responded to therapy. In addition, neither changes in lymphocyte function nor lympho-
toxic for only those eukaryotic cells which express the appropriate cell surface receptor.7,41-42,45-50 Williams et al 42 were able to map the optimal fusion junction by analysis of internal in-frame deletion mutations. The fusion of IL-2 sequences to amino acid 389 of the toxin gave rise to a second generation fusion toxin which was 10-fold more potent then the prototype DAB486IL-2. This observation was recently extended by vanderSpek et al 51 who demonstrated that the maintenance of an intact transmembrane helix 9 is essential for cytotoxic activity (Table 2). The fusion of IL-2 sequences to amino acid 371 of diphtheria toxin, which is positioned in the middle of transmembrane helix 9, resulted in a fusion protein that was ≥ 2,000-fold less cytotoxic, and had a markedly reduced ability to form stable channels in planar lipid bilayers compared to the parental DAB389IL-2. With the solution of the X-ray crystal structure of diphtheria toxin,34,52 it became clear that the fusion of growth factor sequences to amino acid 389 of diphtheria toxin was optimal. Remarkably, amino acid 389 was at the end of a random coil separating the transmembrane from the receptor binding domain of the native toxin.
Clinical evaluation of DAB389IL-2 It has been clear for almost a decade that the high affinity form of the IL-2 receptor is an attractive target for cytotoxic therapy in both cancer and autoimmune disease.53,54 Since the high affinity receptor for IL-2 is transiently expressed on both T and B cells and is not found on other normal tissue, therapeutic agents targeted to this receptor offer the possibility of high selectivity, and as a result correspondingly low adverse effects. In addition, since the diphtheria toxin-based fusion toxins mediate their effect through the ADPribosylation of elongation factor 2, these agents represent a new class of biologic response modifier. A series of phase I/II clinical studies were developed to determine the safety, tolerability, and pharma264
Targeting diphtheria toxin 2R targeted fusion toxin, DAB389IL-2 has been evaluated in the clinic in a focused phase II study in cutaneous T-cell lymphoma. As shown in Table 3, the intravenous administration of DAB389IL-2 to patients with refractory disease has resulted in a remarkable rate of complete and partial remission. In fact, the phase III trial of DAB389IL-2 for this indication has recently been initiated.
cyte subset were detected, and patients were not placed at increased risk of opportunistic infection following treatment. The time course of analysis of DAB486IL-2 concentrations in serum, using a nonlinear mathematical model, showed that the clearance of the fusion toxin followed a one-component model with a t1/2 of approximately 11 minutes at dose levels of 200–400 µg/kg.58 Moreover, the pharmacokinetics of the fusion toxin did not change in a consistent fashion following multiple courses of administration. In many patients increased serum levels of soluble IL-2 receptor (sIL-2R) were detected; however, there was no correlation between the clearance rates of the fusion toxin from circulation, and the level of sIL-2R. Bacha (personal communication) had found previously that the presence of 50,000 units of sIL-2R per ml, failed to inhibit the cytotoxic action of DAB486IL-2 in vitro. Following administration of DAB486IL-2, approximately 60% of the patients had an anamnestic response to the diphtheria toxin component of the fusion protein. Few patients, however, had anti-IL-2 titers prior to study entry. After one or more course of fusion toxin administration approximately half of the patients developed low titers of anti-IL-2 antibodies. As was seen with sIL-2R, the presence of either antidiphtheria toxin-related or anti-IL-2 antibodies did not appear to prevent an anti-tumor response. In fact, the clinical experience with the monoclonal antibody OKT3, and the analysis of neutralizing and nonneutralizing monoclonal antibodies, clearly demonstrated that only those antibodies that effectively blocked binding were neutralizing antibodies.59-61 Since the native diphtheria toxin receptor binding domain is replaced in the construction of the fusion toxins, it was anticipated that pre-existing anti-diphtheria toxoid antibodies would not interfere with the action of DAB486IL-2 in vivo. The results of the phase I/II clinical evaluation of DAB486IL-2 has proved in principal that this fusion toxin is safe, well tolerated, and may induce durable remissions in refractory hematologic malignancies. Based upon these studies, the second generation IL-
Summary and conclusions
Table 3. Refractory cutaneous T-cell lymphoma responses to DAB389IL-2 administration by stage of disease Disease stage I II III IVa
No. patients
Response
% response
8 8 7 9
3CR/3PR 1CR/3PR 1CR/1PR –/–
75 50 29 0
265
A number of years ago we began to ask the fundamental question of whether or not diphtheria toxin could be used as a platform for the development of genetically engineered toxins, in which substitution of the native receptor binding domain with specific growth factors would result in a family of biologically active fusion proteins. These ‘new’ toxins would combine the potent cytotoxic activity of diphtheria toxin with the cell receptor specificity of the growth factor employed as the substitute receptor binding domain. Over the past decade we have learned that the structural genes encoding these fusion toxins could be readily assembled, and that their respective fusion proteins were efficiently expressed and purified in a biologically active conformation from recombinant E. coli. The assembly of these chimeric proteins at the level of the gene has allowed for a detailed analysis of structure function and to an increased understanding of the mechanism by which the catalytic domain of diphtheria toxin is delivered to the target cell cytosol. Most importantly, however, the first of these fusion toxins, DAB389IL-2, has shown remarkable promise as an experimental therapeutic in the treatment of IL-2 receptor positive hematologic malignancies. The intravenous administration of this agent has proved to be safe, well tolerated, and to induce durable remission from disease in a heavily pretreated subset of refractory patients. More recently, this fusion toxin has also been shown to be a safe and effective agent in the treatment of severe psoriasis.62 In fact, the receptor specific action of this fusion toxin has helped to further unravel the underlying basis of psoriasis as an autoimmune, rather than a keratinocyte-based disease. This latter observation opens the possibility that DAB389IL-2 may also be useful in the treatment of those autoimmune diseases in which activated proliferating T cells play a major role in pathogenesis.
J. R. Murphy and J. C. vanderSpek 16. Molloy SS, Thomas L, vanSlyke JK, Stenberg PE, Thomas G (1994) Intracellular trafficking and activation of the furin proprotein convertase: localization to the TGN. EMBO J 13:18-33 17. Moya M, Dautry-Versat A, Goud B, Louvard D, Boquet P (1985) Inhibition of coated pit formation in Hep2 cell blocks the cytotoxicity of diphtheria toxin but not that of ricin toxin. J Cell Biol 101:548-559 18. Morris RE, Gerstein AS, Bonventre PF, Saelinger CB (1985) Receptor-mediated entry of diphtheria toxin into monkey kidney (Vero) cells: electron microscopic evaluation. Infect Immun 50:721-727 19. Fuchs R, Schmid S, Mellman I (1989) A possible role for Na + , K + -ATPase in regulating ATP-dependent endosome acidification. Proc Natl Acad Sci USA 86:539-543 20. Cain CC, Sipe DM, Murphy RF (1989) Regulation of endocytic pH by the Na + K + -ATPase in living cells. Proc Natl Acad Sci USA 86:544-548 21. Kagan BL, Finkelstein A, Colombini M (1981) Diphtheria toxin fragment forms large pores in phospholipid bilayer membranes. Proc Natl Acad Sci USA 78:4950-4954 22. Donovan JJ, Simon MI, Draper RK, Montal M (1981) Diphtheria toxin forms transmembrane channels in planar lipid bilayers. Proc Natl Acad Sci USA 78:172-176 23. Hoch DH, Romero-Mira M, Ehrich BE, Finkelstein A, DasGupta BR, Simpson LL (1985) Channels formed by botulinum, tetanus, and diphtheria toxins in planar lipid bilayers: relevance to translocation of proteins across membranes. Proc Natl Acad Sci USA 82:336-343 24. Papini E, Sandon´a D, Rappuoli R, Montecucco C (1988) On the membrane translocation of diphtheria toxin: at low pH the toxin induces ion channels in cells. EMBO J 7:3353-3359 25. Sandvig K, Olsnes S (1988) Diphtheria toxin-induced channels in Vero cells selective for monovalent cations. J Biol Chem 263:12352-12359 26. O’Keefe DO, Cabiaux V, Choe S, Eisenberg D, Collier RJ (1992) pH-dependent insertion of proteins into membranes: B-chain mutation of diphtheria toxin that inhibits membrane translocation, Glu349 → Lys. Proc Natl Acad Sci USA 89:6202-6206 27. Falnes PO, Madhus IH, Sandvig K, Olsnes S (1992) Replacement of negative by positive charges in the presumed membrane inserted part of diphtheria toxin B fragment: effect on membrane translocation and on formation of channels. J Biol Chem 267:12284-12290 28. vanderSpek J, Cassidy D, Genbauffe F, Huynh P, Murphy JR (1994) An intact transmembrane helix 9 is essential for the efficient delivery of the diphtheria toxin catalytic domain to the cytosol of target cells. J Biol Chem 69:21455-21459 29. Lambotte P, Falmagne P, Capiau C, Zanen J, Ruysschaert J-M, Dirkx J (1980) Primary structure of diphtheria toxin fragment B: structural similarities with lipid-binding domains. J Cell Biol 87:837-840 30. Stenmark H, Ariansen S, Afanasieu BN, Olsnes S (1992) Interactions of diphtheria toxin B fragment with cells. Role of amino and carboxy-terminal regions. J Biol Chem 281:619-625 31. vanderSpek JC, Mindel J, Finkelstein A, Murphy JR (1993) Structure function analysis of the transmembrane domain of the interleukin-2 receptor target fusion toxin DAB389IL-2: The amphipathic helical region of the transmembrane domain is essential for the efficient delivery of the catalytic domain to the cytosol of target cells. J Biol Chem 268:12077-12082 32. Wiedlocha A, Madhus IH, Mach H, Middaugh CR, Olsnes S (1992) Tight folding of acidic fibroblast growth factor prevents its translocation into the cytosol with diphtheria toxin as vector. EMBO J 11:4835-4842 33. Falnes PO, Choe S, Madhus IH, Wilson BA, Olsnes S (1994) Inhibition of membrane translocation of diphtheria toxin
Acknowledgements We thank Larry Cosenza for the ribbon diagram of native diphtheria toxin. JRM and JCvdS are partially supported by grant CA-60934 from the National Cancer Institute. JvdS is a Special Fellow of the Leukemia Society of America.
References 1. Uchida T, Gill DM, Pappenheimer AM, Jr (1971) Mutation in the structural gene for diphtheria toxin carried by temperate phage β. Nature 233:8-11 2. Murphy JR, Pappenheimer AM Jr, Tayart de Borms S (1974) Synthesis of diphtheria tox gene products in Escherichia coli extracts. Proc Natl Acad Sci USA 71:11-15 3. Boquet P, Silverman MS, Pappenheimer AM Jr, Vernon WB (1976) Binding of triton X-100 to diphtheria toxin, crossreacting material 45, and their fragments. Proc Natl Acad Sci USA 73:4449-4453 4. Middlebrook JL, Dorland RB, Leppla SH (1978) Association of diphtheria toxin with Vero cells: demonstration of a receptor. J Biol Chem 253:7325-7330 5. Uchida T, Pappenheimer AM Jr, Greaney R (1973) Diphtheria toxin and related proteins: isolation and properties of mutant proteins serologically related to diphtheria toxin. J Biol Chem 248:3838-3844 6. Hayakawa S, Uchida T, Mekada E, Moynihan MR, Okada Y (1983) Monoclonal antibody against diphtheria toxin: effect on toxin binding and entry into cells. J Biol Chem 258:4311-4317 7. Murphy JR, Bishai W, Borowski M, Miyanohara A, Boyd J, Nagle S (1986) Genetic construction, expression, and melanomaselective cytotoxicity of a diphtheria toxin α-melanocyte stimulating hormone fusion protein. Proc Natl Acad Sci USA 83:8258-8262 8. Greenfield L, Johnson VG, Youle RJ (1987) Mutations in diphtheria toxin separate binding from entry and amplify immunotoxin selectivity. Science 238:536-539 9. Myers DA, Villemez CL (1988) Specific chemical cleavage of diphtheria toxin with hydroxyamine: purification and characterization of the modified proteins. J Biol Chem 263:17122-17127 10. Rolf JM, Eidels L (1993) Structure-function analyses of diphtheria toxin by use of monoclonal antibodies. Infect Immun 61:944-1003 11. Schalken JA, Roebuck AJM, Oomen PPCA, Wagenaar SS, Debruyne FM, Bloemers HPJ, van de Ven WJM (1987) Furin gene expression as a discriminating marker for small cell and non small cell lung carcinomas. J Clin Invest 8:1545-1549 12. Barr PJ (1991) Mammalian subtilisins: the long sought dibasic processing endoproteases. Cell 66:1-3 13. Williams DP, Wen Z, Watson RS, Boyd J, Strom TB, Murphy JR (1990) Cellular processing of the fusion toxin DAB486IL-2 and efficient delivery of diphtheria toxin fragment A to the cytosol of target cells requires Arg194. J Biol Chem 265:20673-20677 14. Klimpel KR, Molloy SS, Thomas G, Leppla S (1992) Anthrax toxin protective antigen is activated by a cell surface protease with the sequence specificity and catalytic properties of furin. Proc Natl Acad Sci USA 89:10277-10281 15. Tsuneoka M, Nakagama K, Hatsuzawa K, Komada M, Kitamura N, Mekada E (1993) Evidence for involvement of furin in cleavage and activation of diphtheria toxin. J Biol Chem 268:26461-26465
266
Targeting diphtheria toxin
34. 35. 36.
37.
38. 39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
A-fragment by internal disulfide bridges. J Biol Chem 269:8402-8407 Choe S, Bennett MJ, Fujii G, Curmi PMG, Kantardjieff KA, Collier RJ, Eisenberg D (1992) The crystal structure of diphtheria toxin. Nature 357:216-222 Moskaug JO, Sandvig K, Olsnes S (1987) Cell-mediated reduction of the interfragment disulfide in nicked diphtheria toxin. J Biol Chem 262:10339-10345 Papini E, Rappuoli R, Muriga M, Montecucco C (1993) Cell penetration of diphtheria toxin. Reduction of the interchain disulfide bridge is the rate-limiting step of translocation into the cytosol. J Biol Chem 268:1567-1574 Mandel R, Ryser H-P, Ghani F, Wu M, Peak D (1993) Inhibition of a reductive function of the plasma membrane by bacitracin and antibodies against protein disulfide isomerase. Proc Natl Acad Sci USA 90:4112-4116 Freedman RB (1989) Protein disulfide isomerase multiple role in the modification of nascent secretory proteins. Cell 57:1069-1072 Bacha P, Murphy JR, Reichlin S (1983) Thyrotropin-releasing hormone-diphtheria toxin-related polypeptide conjugates: potential role of the hydrophobic domain in toxin entry. J Biol Chem 258: 1565-1570 Cawley DB, Herschman HR, Gilliland DG, Collier RJ (1980) Epidermal growth factor-toxin A chain conjugates: EGF-ricin A is a potent toxin while EGF-diphtheria fragment A is nontoxic. Cell 22:563-570 Williams D, Parker K, Bishai W, Borowski M, Genbauffe F, Strom TB, Murphy JR (1987) Diphtheria-toxin receptor binding domain substitution with interleukin-2: genetic construction and properties of a diphtheria toxin-related interleukin-2 fusion protein. Protein Engn 1:493-498 Williams D, Snider CE, Strom TB, Murphy JR (1990) Structure function analysis of IL-2 toxin (DAB486IL-2): fragment B sequences required for the delivery of fragment A to the cytosol of target cells. J Biol Chem 265:11885-11889 Bacha P, Waters C, Williams J, Murphy JR, Strom TB (1988) Interleukin-2 targeted cytotoxicity: selective action of a diphtheria toxin-related interleukin-2 fusion protein. J Exp Med 167:612-622 Waters CA, Schimke P, Snider CE, Itoh K, Smith KA, Nichols JC, Strom TB, Murphy JR (1990) Interleukin-2 receptor targeted cytotoxicity: receptor binding requirements for entry of IL2-toxin into cells. Eur J Immunol 20:785-791 Aullo P, Alcani J, Popoff MR, Klatzman DR, Murphy JR, Boquet P (1992) In vitro effects of a recombinant diphtheria-human CD4 fusion toxin on acute and chronically HIV-1 infected cells. EMBO J 12:921-931 Lakkis F, Steele A, Pacheco-Silva A, Kelley VE, Strom TB, Murphy JR (1991) Interleukin-2 receptor targeted cytotoxicity: genetic construction and properties of diphtheria toxin-related interleukin-4 fusion toxins. Eur J Immunol 21:2253-2258 Wen Z, Tao X, Lakkis F, Kiyokawa T, Murphy JR (1991) Expression, purification, and α-melanocyte stimulating hormone receptor-specific toxicity of DAB-α-MSH fusion toxins. J Biol Chem 266:12289-12293 Shaw JP, Akiyoshi DE, Arrigo DA, Rhoad AE, Sullivan B, Thomas J, Genbauffe FS, Bacha P, Nichols JC (1991) Cytotoxic properties of DAB486EGF and DAB389EGF, epidermal growth factor (EGF) receptor-targeted fusion toxins. J Biol Chem 266:13449-13455
49. Jean L-F, Murphy JR (1992) Diphtheria toxin receptor binding domain substitution with interleukin-6: genetic construction and interleukin-6 receptor specific action of a diphtheria toxinrelated interleukin-6 fusion protein. Protein Engn 4:989-994 50. Sweeney EB, vanderSpek JC, Foss F, Murphy JR (1995) Genetic construction and characterization of DAB389IL-7: A novel agent for the elimination of IL-7 receptor positive cells. (submitted) 51. vanderSpek J, Howland K, Friedman T, Murphy JR (1994) Maintenance of the hydrophobic face of the diphtheria toxin amphipathic transmembrane helix 1 is essential for the efficient delivery of the catalytic domain to the cytosol of target cells. Protein Engn 7:985-989 52. Bennett MJ, Choe S, Eisenberg D (1994) Domain swapping: entangling alliances between proteins. Proc Natl Acad Sci USA 91:3127-3131 53. Waldmann TA (1986) The structure, function, and expression of interleukin-2 receptors on normal and malignant T cells. Science 232:727-732 54. Waldmann TA (1990) The multichain interleukin-2 receptor. A target for immunotherapy in lymphoma, autoimmune disorders, and organ allografts. J Am Med Assoc 263:272-274 55. LeMaistre CF, Meneghetti C, Rosenblum M, Reuben J, Parker K, Shaw J, Woodworth T, Parkinson D (1992) Phase I trial of an interleukin-2 receptor (IL-2R) fusion toxin (DAB486IL-2) in hematologic malignancies expressing the IL-2 receptor. Blood 79:2547-2554 56. Schwartz G, Tepler I, Charette J, Kadin L, Parker K, Woodworth T, Schnipper L (1992) Complete response of a Hodgkin’s lymphoma in a phase I trail of DAB486IL-2. Blood 79:175a 57. Hesketh P, Caguioa P, Koh H, Dewey H, Facada A, McCaffrey R, Parker K, Nylen P, Woodworth T (1993) Clinical activity of a cytotoxic fusion protein in the treatment of cutaneous T cell lymphoma. J Clin Oncol 11:1628-1690 58. LeMaistre CF, Craig FE, Meneghetti C, McMullin B, Parker K, Reuben J, Boldt DH, Rosenblum M, Woodworth T (1993) Phase I trail of a 90-minute infusion of the fusion toxin DAB486IL-2 in hematologic cancers. Cancer Res 53:3930-3934 59. Chatenoud L, Jonker M, Villemain F, Goldstein G, Bach JF (1986) The human immune response to the OKT3 monoclonal antibody is oligoclonal. Science 232:1406-1408 60. Bacha P, Forte SE, Perper SJ, Trentham DE, Nichols JC (1992) Anti-arthritic effects demonstrated by an interleukin-2 receptortargeted cytotoxin (DAB486IL-2) in rat adjuvant arthritis. Eur J Immunol 22:1673-1679 61. Zucker DR, Murphy JR (1984) Monoclonal antibody analysis of diphtheria toxin. I. Localization of epitopes and neutralization of cytotoxicity. Molec Immunol 21:785-793 62. Gottlieb SL, Gilleaudeau P, Johnson R, Estis L, Woodworth T, Gottlieb AB, Krueger JG (1995) Response of psoriasis to a lymphocyte-selective toxin (DAB389IL-2) suggests a primary immune, but not keratinocyte, pathogenic basis. Nature Med 1:442-447 63. Kraulis PJ (1991) MOLESCRIPT: a program to produce both detailed and schematic plots of protein structures. J Appl Crystallogr 24:946-950 64. Kiyokawa T, Williams DP, Snider CE, Strom TB, Murphy JR (1991) Protein engineering of diphtheria toxin-related interleukin-2 fusion toxins to increase biologic potency for high affinity interleukin-2 receptor bearing cells. Protein Engn 4:463-468
267
CANCER BIOLOGY, Vol 6, 1995: pp 259–267
Targeting diphtheria toxin to growth factor receptors John R. Murphy and Johanna C. vanderSpek
negative host. The second experiment addressed the possibility of creating ‘new’ toxins that would be selectively active and eliminate only those cells expressing target receptor on their surface. Although these experiments were approved by the RAC, this research was put on administrative hold for almost four years, in part, because of its controversial nature. By the fall of 1985, we were finally allowed to proceed with only those experiments involving the genetic substitution of the native diphtheria toxin receptor binding domain with α-MSH at the National Institutes of Allergy and Infectious Diseases (NIAID) BL-4 facility at Frederick, Maryland.
Biochemical, genetic and X-ray crystallographic analysis of diphtheria toxin have demonstrated that the native toxin is composed of three structural domains that function in an ordered fashion to intoxicate a eukaryotic cell. With the knowledge that, if delivered to the cytosol, a single molecule of the catalytic domain is lethal for the cell, we have used recombinant DNA methods to genetically replace the native toxin receptor binding domain with a series of growth factors. The resulting diphtheria toxin-related cytokine fusion proteins, or fusion toxins bind to their respective receptors, are internalized by receptor-mediated endocytosis, and efficiently eliminate target cell populations by the adenosine diphosphate ribosylation of elongation factor 2. Based upon the results of preclinical studies, DAB486IL-2, DAB389IL-2 and DAB389EGF have, or are in the process of being evaluated in Phase I/II clinical trials. To date, administration of the diphtheria toxin-based fusion proteins targeted toward the high affinity IL-2 receptor have been found to be safe, well tolerated, and capable of inducing remission in refractory hematologic malignancies.
The diphtheria toxin platform for genetic engineering The choice of diphtheria toxin as the toxophore for receptor binding domain substitution was based upon a number of earlier observations. In their classic study, Uchida, Gill and Pappenheimer1 demonstrated that the structural gene for diphtheria toxin was carried by corynebacteriophage β. This study also provided the foundation for subsequent studies on the structure function relationships of diphtheria toxin by demonstrating that the enzymatically active A fragment was positioned on the N-terminal end of the toxin; whereas, the receptor binding domain of the toxin was carried on fragment B. Shortly thereafter, Murphy et al 2 used β-phage DNA to program S-30 extracts of E. coli and demonstrated, in this coupled transcription translation system, that biologically active diphtheria toxin could be synthesized in vitro. Using [3H]-Triton X-100, Boquet et al 3 clearly demonstrated that fragment B of both diphtheria toxin and the non-toxic mutant CRM45 carried a hydrophobic domain which had properties of integral membrane proteins. These investigators postulated that the role of this hydrophobic domain was to facilitate the delivery of fragment A across the eukaryotic cell membrane and into the cytosol. Thus, it was known quite early that native diphtheria toxin was a three domain protein consisting of the enzymatically active domain (fragment A), the hydrophobic
Key words: cytotoxicity / diphtheria toxin / growth factor / psoriasis / receptor ©1995 Academic Press Ltd
ALMOST 15 years ago we submitted a proposal to the Recombinant DNA Advisory Committee (RAC) to conduct two experiments under Biosafety Level 4 (BL4). The first experiment was to clone the structural gene for diphtheria toxin in Escherichia coli K-12 and, based upon the assumption that biologically active diphtheria toxin would be expressed and secreted into the periplasmic space, the second experiment proposed the genetic substitution of the native diphtheria toxin receptor binding domain with the polypeptide hormone α-melanocyte stimulating hormone (α-MSH). The initial experiment was designed to test the feasibility of expressing a potent toxin from a Gram positive organism in a heterologous Gram From Evans Department of Clinical Research and Department of Medicine, Boston University Medical Center Hospital, Boston, MA 02118, USA ©1995 Academic Press Ltd 1044-579X/95/050259 + 09 $12.00/0
259
J. R. Murphy and J. C. vanderSpek subsequent to receptor binding by cell surface associated proteases. In either case, it is clear that processing must occur before the catalytic domain is translocated to the cytosol. The protease recognition sequence of diphtheria toxin is Arg190-Val191Arg192-Arg193 which has been shown to be a site for the endoprotease furin.11,12 Additional support for a furin-mediated cleavage of the toxin as an essential step in the intoxication process comes from a sitedirected mutational analysis of DAB486IL-2, in which destruction of the furin recognition site resulted in a marked loss of cytotoxic activity in the fusion toxin.13 More recently, it has been shown that furin can be used to specifically nick intact diphtheria toxin in vitro.14 Tsuneoka et al 15 have shown that LoVo cells, which do not produce furin are not sensitive to the action of diphtheria toxin; however, upon transfection of LoVo cells with the gene encoding furin, the cells become sensitive to the toxin. Since it has been shown that furin cycles from the trans-Golgi to the membrane,16 it is almost certain that intact diphtheria toxin, and the diphtheria-related fusion toxins are processed on the cell surface by furin.
domain (N-terminal portion of fragment B), and the receptor binding domain (C-terminal portion of fragment B). The process by which diphtheria toxin intoxicates sensitive eukaryotic cells involves at least the following steps: (i) binding of the toxin to its cell surface receptor, (ii) activation of the catalytic domain by a proteolytic cleavage (‘nicking’) of the toxin in a sensitive exposed 14 amino acid loop that is subtended by Cys186 and Cys201, (iii) internalization of the bound toxin into endosomes by receptor-mediated endocytosis, and following acidification of the endocytic vesicle, (iv) the facilitated delivery of the catalytic domain across the endocytic vesicle membrane and into the cytosol. Once delivered to the cytosol fragment A rapidly catalyses the adenosine diphosphate ribosylation of elongation factor 2 which results in the inhibition of protein synthesis and subsequent death of the cell.
Receptor binding domain The first step in the intoxication process is the specific binding of diphtheria toxin to its cell surface receptor. Middlebrook et al 4 were able to correlate the apparent sensitivity of a given cell line to diphtheria toxin with the number of receptors on the cell surface. The initial localization of the diphtheria toxin receptor binding domain to the carboxy-terminal region of fragment B was based upon the findings that CRM45, a pre-mature chain termination mutant of the toxin which lacked the C-terminal 15,000 dalton region, failed to block the toxic activity of diphtheria toxin on cells.1,5 In addition, observations made by many investigators using different approaches strongly suggested that the functional native receptor binding domain was positioned in the carboxy-terminal 50 amino acids of the toxin.6-10
Transmembrane domain It has been known for over a decade that once diphtheria toxin is bound to its cell surface receptor, it is internalized into the cell by receptor-mediated endocytosis.17,18 Early endocytic vesicles are known to be acidified by specific vesicular ATPases to an average pH value of 6.2.19,20 It is well known that diphtheria toxin must pass through an acidic compartment in order to deliver the catalytic domain to the cytosol. Over a decade ago, it was recognized that under acidic conditions diphtheria toxin and CRM45 will spontaneously insert into the plane of lipid bilayers and form channels.21,22 Moreover, the diameter of the channel has been reported to be 18 A˚ 21,23 which is large enough for a denatured, extended fragment A to pass through and be delivered to the cytosol. Importantly, diphtheria toxin-induced channels have been observed in both Vero and CHO cell membranes following a low pH-pulse.24,25 The channel formation that has been observed in both artificial and cellular membranes involves the insertion of transmembrane domain α-helices. O’Keefe et al 26 found that the introduction of the E349K mutation, which is positioned in the short loop connecting the transmembrane helices 8 and 9, resulted in both a decrease in the cytotoxic potency and channel-forming activity of the mutant toxin.
Processing/activation domain Once bound to its cell surface receptor, intact diphtheria toxin must be processed into an active two chain protein through proteolytic processing, or nicking. Intact diphtheria toxin has an exposed serine-protease sensitive 14-amino acid loop that is subtended by a disulfide bond between Cys185 and Cys201. As shown in Figure 1, the cleavage domain of the toxin is positioned between the catalytic and transmembrane domains. Nicking, or processing of the toxin may occur either prior to the toxin binding to its cell surface receptor by serum proteases or 260
Targeting diphtheria toxin These investigators proposed that following protonation of E349, and possibly D325 in the endocytic vesicle, transmembrane helices 8 and 9 would spontaneously insert into the plane of the membrane. Further, upon exposure of the connecting loop to the
neutral pH environment of the cytosol, E349 would become deprotonated and its charge would lock, or anchor this α-helical hairpin in the membrane. In an analogous fashion, Falnes et al 27 reported that the diphtheria toxin-related site-directed mutant D318K
Figure 1. Ribbon diagram of the X-ray crystal structure of native diphtheria toxin34 as modified by Bennett et al.52 The catalytic domain (red), transmembrane domain (magenta), and receptor binding domain (blue) are shown. N, N-terminal end of the toxin; PSL, 14 amino-acid protease sensitive loop which separates the catalytic from transmembrane and receptor binding domains; C, C-terminal end of the toxin. The ribbon diagram was generated using MOLESCRIPT.63
261
J. R. Murphy and J. C. vanderSpek complete unfolding of the catalytic domain comes from the observations of Falnes et al. 33 Using the X-ray crystal structure of diphtheria toxin,34 Falnes et al 33 showed the introduction of paired cysteine residues in the catalytic domain, which were likely to form disulfide bridges, resulted in the formation of mutants with decreased cytotoxic activity. While it is clear that the disulfide bond between the catalytic and transmembrane domains (Cys186– Cys201) must be reduced in order to release the catalytic domain, the precise cellular location of where this reduction occurs is unknown. Moskaug et al 35 provided evidence that the reduction of the disulfide bond occurs either when the transmembrane domain becomes buried in the lipid core or when the bond is exposed to the cytosol. More recently, Papini et al 36 demonstrated that the reduction of the disulfide bond occurs after the low-pH insertion of the transmembrane domain into an early endosomal compartment. Since membrane-impermeant sulfhydryl blockers inhibit the action of diphtheria toxin, it has been proposed that the reduction of the disulfide bond between the catalytic and transmembrane domains could happen either at the level of the cell surface or in the lumen of the endocytic vesicle. In support of this hypothesis, Mandel et al 37 have provided evidence that protein disulfide isomerase (PDI) plays a major role in the diphtheria toxin mediated intoxication of Vero cells. Since PDI has been reported to be primarily localized in the endoplasmic reticulum38 and direct evidence demonstrating that PDI plays a role in the intoxication process is lacking, this hypothesis awaits further experimental support.
[D318 is positioned in the loop connecting transmembrane α-helices 5 and 6] also had reduced cytotoxic potency and impaired ability to form channels in artificial membrane bilayers. vanderSpek et al 28 have shown that an intact transmembrane helix 9 is also required for the selective toxicity DAB389IL-2. Inframe deletion mutants that lack the C-terminal portion of transmembrane helix 9 had markedly reduced cytotoxic potency and were incapable of forming stable channels in artificial membranes. The amino-terminal region of the transmembrane domain is amphipathic and has been shown to be structurally homologous to apolipoprotein A1.29 The introduction of mutations into this portion of diphtheria toxin have been shown to result in mutants with a decreased ability to bind to the diphtheria toxin receptor.30 In contrast, mutations in this region of the fusion toxin molecule do not effect receptor binding affinity; however, they may result in non-toxic mutants in which neither receptor binding activity nor channel formation are impaired.31 Since these mutants bind with high affinity and are capable of forming channels in artificial membranes, they define a hitherto unknown step in the intoxication process.
Catalytic domain Relatively little is known of the molecular details involving the translocation of the catalytic domain of diphtheria toxin through the endocytic vesicle membrane and into the cytosol. Moreover, most of our current understanding of the mechanism by which the catalytic domain enters the cytosol is based upon indirect observations. For example, evidence supporting the hypothesis that the catalytic domain must completely unfold comes from the early observations that the pore formed by insertion of the transmembrane domain into the membrane is approximately 18 A˚ in diameter; only large enough to allow the passage of a completely unfolded catalytic domain. More recently Wiedlocha et al 32 described the construction of a fusion protein in which acidic fibroblast growth factor (aFGF) was linked to the N-terminus of the catalytic domain. While this fusion protein was as toxic toward Vero cells as the native toxin, the addition of heparin to the assay medium (i.e. heparin induces a tight folding of aFGF) resulted in a complete loss of toxicity. These investigators reasoned that the heparin-induced tight folding of the aFGF component of the fusion protein caused a block in the translocation of the catalytic domain into the cytosol. Perhaps the most compelling evidence in support of
Diphtheria toxin-related growth factor fusion proteins In 1980, we39 began to investigate the role that the hydrophobic domain of fragment B played in the delivery of fragment A to the cytosol of target cells. This study was prompted by the observation that an epidermal growth factor (EGF) ricin A-chain conjugate toxin was exquisitely cytotoxic for cells bearing the EGF receptor; whereas, the analogous conjugate toxin assembled with fragment A of diphtheria toxin was essentially devoid of activity.40 We reasoned that the failure of diphtheria toxin fragment A-based conjugate to be active was likely due to either the lack of a hydrophobic domain that facilitated the entry of the catalytic domain to the cytosol or to stearic 262
Targeting diphtheria toxin of the fusion toxins might be useful in the treatment of human disease. Murphy et al 7 described the construction and properties of a fusion protein, DAB486α-MSH, in which the native diphtheria toxin receptor binding domain was replaced with α-melanocyte stimulating hormone (α-MSH). This construct employed a unique SphI restriction endonuclease site in the diphtheria toxin structural gene such that amino acid 486 served as the fusion junction between diphtheria toxin-related sequences and α-MSH. Despite the fact that DAB486α-MSH was subject to marked proteolytic degradation in E. coli, we were able to partially purify sufficient levels of the fusion toxin to demonstrate α-MSH receptor specific toxicity. Importantly, we were able to demonstrate by guinea pig and mouse challenge experiments conducted under BL-4 containment, that the recombinant E. coli expressing DAB486α-MSH were avirulent. This data supported the argument that further construction of novel recombinant diphtheria toxin-related cytokine fusion genes could safely be conducted in our laboratory in Boston under BL-3 containment. Since DAB486α-MSH was subject to marked proteolytic degradation, and the protease(s) sensitive sites appeared to be close to the fusion junction between diphtheria toxin and α-MSH, we reasoned that a fusion toxin constructed with a growth factor of larger mass might provide stearic hinderance and thereby minimize degradation. In fact this was found to be the case, and much of our current understanding of the diphtheria toxin-related fusion toxins comes from DAB486IL-2 and DAB389IL-2 in which the native receptor binding domain of the toxin was replaced with interleukin 2.41-44 As shown in Table 1, a variety of growth factors have been used to replace the native receptor binding domain of the toxin. In all instances, these fusion toxins have been shown to be selectively
constraints imposed by the EGF portion of the conjugate. Accordingly, we modified the imidazole ring of the central histidine of thyrotropin releasing hormone (TRH; pyro-Glu-His-Pro-NH2) by the addition of acetylcystamine. Once modified, acetylcystaminyl-THR was then coupled through a disulfide bond to either CRM26 (fragment A) or CRM45. While both CRM26-TRH and CRM45-TRH were found to specifically bind to the TRH receptor on CH3 rat pituitary cells, only the CRM45-TRH conjugate was cytotoxic (IC50 = 3 3 10–9M). In marked contrast, CRM26-TRH was devoid of activity at concentrations greater than 10–7 M. These experiments demonstrated that the hydrophobic domain of fragment B was essential in order to facilitate the delivery of fragment A to the cytosol. The most troubling aspect encountered in the assembly of the CRM45-TRH conjugates was that the specific toxicity (IC50/mg conjugate) varied widely from preparation to preparation. This variation was presumably due to differences in the extent of acetylcystaminylation of native TRH, differences in either the extent of modification of CRM45 with N-succinimidyl-3-(2-pyridyldithio)-proprionate (SPDP), and/or in the extent of coupling of the modified proteins into the hormone toxin conjugate. In order to overcome the uncertainties associated with chemical conjugation, we turned to the methodologies of protein engineering and recombinant DNA to assemble fusion genes in which the native receptor binding domain of diphtheria toxin was replaced with specific ligands. From the outset, the prospect of using recombinant DNA methods to assemble the structural genes encoding bacterial toxin growth factor fusion proteins, or fusion toxins, offered significant advantages over chemical conjugation in the assembly of chimeric proteins. Most importantly, the fusion junction, or point at which the substitute receptor binding domain was linked to the toxin fragment, was precisely determined. Subsequent expression in recombinant E. coli then resulted in the synthesis of a single gene product rather than the mixture of isomeric forms that result from chemical conjugation of two proteins. Thus, we reasoned that by in-frame deletion analysis, it should be possible to determine the diphtheria toxin fragment B (transmembrane domain) structure required for the efficient delivery of fragment A through the membrane and into the cytosol. Furthermore, with the proviso that the distribution of the targeted receptor was limited to subsets of cells that were involved in a pathogenic process, we envisioned that at least some
Table 1. Diphtheria toxin-based cell receptor-targeted fusion proteins Fusion toxin DAB486α-MSH DAB389α-MSH DAB486IL-2 DAB389IL-2 DAB389IL-4 DAB389IL-6 DAB389IL-7 DAB389EGF DAB486CD4
263
Receptor
Cytotoxicity (IC50)
Ref
α-MSH α-MSH IL-2 IL-2 IL-4 IL-6 IL-7 EGF HIV gp120
n.d. 3×10–11M 1×10–11M 1×10–12M 2×10–10M 2×10–11M 1×10–10M 1×10–11M 1×10–9M
7 47 41 64 46 49 50 48 45
J. R. Murphy and J. C. vanderSpek Table 2. Cytotoxicity, binding affinity, and planar lipid bilayer conductance of DAB389IL-2 and in-frame deletion mutants (modified from ref 51) Fusion toxin DAB389IL-2 DAB (∆381–387)389IL-2 DAB (∆369–387)389IL-2
Cytotoxicity (IC50×10–12M)
Binding affinity (Kd×10–12M)
Conductance (pS)
3 70 6,000
6.6 8.0 9.1
42 40 *33*
*Unstable channels.
cokinetics of DAB486IL-2 in patients with refractory hematologic malignancies.55-57 These early studies employed DAB486IL-2 as the experimental therapeutic in order to establish the ‘proof of principal’ that the diphtheria toxin-based fusion toxins would show some degree of efficacy in the treatment of human disease. In fact, several of these clinical studies were initiated prior to the isolation and characterization of the more potent DAB389IL-2 form of the fusion toxin. In order to be eligible for study entry, patients had to present with a hematologic malignancy that was refractory to standard treatment regimens. Patients of either sex over the age of 17 years could be treated if they had adequate hepatic and renal function and a Karnofsky performance score of 70% or greater. The initial clinical trials were designed as a three patient cohort dose escalation in which single and multiple doses of the fusion toxin were administered by intravenous injection, either as a bolus, or 90-minute infusion. The initial dose was 700 ng/kg per day and was escalated to 400 µg/kg per day. Patients were monitored for adverse effects and response and were scored according to National Cancer Institute criteria. Patients with no evidence of disease for at least four weeks were classified as complete responders (CR); patients whose tumor burden decreased by ≥ 50% for at least four weeks were classified as partial responders (PR); and patients whose tumor burden decreased by ≥ 50% were classified as minor responders. The intravenous administration of DAB486IL-2 was found to be well tolerated at all dose levels. The adverse effects were generally mild and included nausea/vomiting, hypersensitivity, fever/malaise/ chills, and elevations in serum hepatic transminases. Renal insufficiency defined the maximum tolerated dose at levels of DAB486IL-2 above 400 µg/kg per day. Importantly, in all instances the adverse effects were transient, not cumulative, and did not preclude repeated administration of the fusion toxin to patients who responded to therapy. In addition, neither changes in lymphocyte function nor lympho-
toxic for only those eukaryotic cells which express the appropriate cell surface receptor.7,41-42,45-50 Williams et al 42 were able to map the optimal fusion junction by analysis of internal in-frame deletion mutations. The fusion of IL-2 sequences to amino acid 389 of the toxin gave rise to a second generation fusion toxin which was 10-fold more potent then the prototype DAB486IL-2. This observation was recently extended by vanderSpek et al 51 who demonstrated that the maintenance of an intact transmembrane helix 9 is essential for cytotoxic activity (Table 2). The fusion of IL-2 sequences to amino acid 371 of diphtheria toxin, which is positioned in the middle of transmembrane helix 9, resulted in a fusion protein that was ≥ 2,000-fold less cytotoxic, and had a markedly reduced ability to form stable channels in planar lipid bilayers compared to the parental DAB389IL-2. With the solution of the X-ray crystal structure of diphtheria toxin,34,52 it became clear that the fusion of growth factor sequences to amino acid 389 of diphtheria toxin was optimal. Remarkably, amino acid 389 was at the end of a random coil separating the transmembrane from the receptor binding domain of the native toxin.
Clinical evaluation of DAB389IL-2 It has been clear for almost a decade that the high affinity form of the IL-2 receptor is an attractive target for cytotoxic therapy in both cancer and autoimmune disease.53,54 Since the high affinity receptor for IL-2 is transiently expressed on both T and B cells and is not found on other normal tissue, therapeutic agents targeted to this receptor offer the possibility of high selectivity, and as a result correspondingly low adverse effects. In addition, since the diphtheria toxin-based fusion toxins mediate their effect through the ADPribosylation of elongation factor 2, these agents represent a new class of biologic response modifier. A series of phase I/II clinical studies were developed to determine the safety, tolerability, and pharma264
Targeting diphtheria toxin 2R targeted fusion toxin, DAB389IL-2 has been evaluated in the clinic in a focused phase II study in cutaneous T-cell lymphoma. As shown in Table 3, the intravenous administration of DAB389IL-2 to patients with refractory disease has resulted in a remarkable rate of complete and partial remission. In fact, the phase III trial of DAB389IL-2 for this indication has recently been initiated.
cyte subset were detected, and patients were not placed at increased risk of opportunistic infection following treatment. The time course of analysis of DAB486IL-2 concentrations in serum, using a nonlinear mathematical model, showed that the clearance of the fusion toxin followed a one-component model with a t1/2 of approximately 11 minutes at dose levels of 200–400 µg/kg.58 Moreover, the pharmacokinetics of the fusion toxin did not change in a consistent fashion following multiple courses of administration. In many patients increased serum levels of soluble IL-2 receptor (sIL-2R) were detected; however, there was no correlation between the clearance rates of the fusion toxin from circulation, and the level of sIL-2R. Bacha (personal communication) had found previously that the presence of 50,000 units of sIL-2R per ml, failed to inhibit the cytotoxic action of DAB486IL-2 in vitro. Following administration of DAB486IL-2, approximately 60% of the patients had an anamnestic response to the diphtheria toxin component of the fusion protein. Few patients, however, had anti-IL-2 titers prior to study entry. After one or more course of fusion toxin administration approximately half of the patients developed low titers of anti-IL-2 antibodies. As was seen with sIL-2R, the presence of either antidiphtheria toxin-related or anti-IL-2 antibodies did not appear to prevent an anti-tumor response. In fact, the clinical experience with the monoclonal antibody OKT3, and the analysis of neutralizing and nonneutralizing monoclonal antibodies, clearly demonstrated that only those antibodies that effectively blocked binding were neutralizing antibodies.59-61 Since the native diphtheria toxin receptor binding domain is replaced in the construction of the fusion toxins, it was anticipated that pre-existing anti-diphtheria toxoid antibodies would not interfere with the action of DAB486IL-2 in vivo. The results of the phase I/II clinical evaluation of DAB486IL-2 has proved in principal that this fusion toxin is safe, well tolerated, and may induce durable remissions in refractory hematologic malignancies. Based upon these studies, the second generation IL-
Summary and conclusions
Table 3. Refractory cutaneous T-cell lymphoma responses to DAB389IL-2 administration by stage of disease Disease stage I II III IVa
No. patients
Response
% response
8 8 7 9
3CR/3PR 1CR/3PR 1CR/1PR –/–
75 50 29 0
265
A number of years ago we began to ask the fundamental question of whether or not diphtheria toxin could be used as a platform for the development of genetically engineered toxins, in which substitution of the native receptor binding domain with specific growth factors would result in a family of biologically active fusion proteins. These ‘new’ toxins would combine the potent cytotoxic activity of diphtheria toxin with the cell receptor specificity of the growth factor employed as the substitute receptor binding domain. Over the past decade we have learned that the structural genes encoding these fusion toxins could be readily assembled, and that their respective fusion proteins were efficiently expressed and purified in a biologically active conformation from recombinant E. coli. The assembly of these chimeric proteins at the level of the gene has allowed for a detailed analysis of structure function and to an increased understanding of the mechanism by which the catalytic domain of diphtheria toxin is delivered to the target cell cytosol. Most importantly, however, the first of these fusion toxins, DAB389IL-2, has shown remarkable promise as an experimental therapeutic in the treatment of IL-2 receptor positive hematologic malignancies. The intravenous administration of this agent has proved to be safe, well tolerated, and to induce durable remission from disease in a heavily pretreated subset of refractory patients. More recently, this fusion toxin has also been shown to be a safe and effective agent in the treatment of severe psoriasis.62 In fact, the receptor specific action of this fusion toxin has helped to further unravel the underlying basis of psoriasis as an autoimmune, rather than a keratinocyte-based disease. This latter observation opens the possibility that DAB389IL-2 may also be useful in the treatment of those autoimmune diseases in which activated proliferating T cells play a major role in pathogenesis.
J. R. Murphy and J. C. vanderSpek 16. Molloy SS, Thomas L, vanSlyke JK, Stenberg PE, Thomas G (1994) Intracellular trafficking and activation of the furin proprotein convertase: localization to the TGN. EMBO J 13:18-33 17. Moya M, Dautry-Versat A, Goud B, Louvard D, Boquet P (1985) Inhibition of coated pit formation in Hep2 cell blocks the cytotoxicity of diphtheria toxin but not that of ricin toxin. J Cell Biol 101:548-559 18. Morris RE, Gerstein AS, Bonventre PF, Saelinger CB (1985) Receptor-mediated entry of diphtheria toxin into monkey kidney (Vero) cells: electron microscopic evaluation. Infect Immun 50:721-727 19. Fuchs R, Schmid S, Mellman I (1989) A possible role for Na + , K + -ATPase in regulating ATP-dependent endosome acidification. Proc Natl Acad Sci USA 86:539-543 20. Cain CC, Sipe DM, Murphy RF (1989) Regulation of endocytic pH by the Na + K + -ATPase in living cells. Proc Natl Acad Sci USA 86:544-548 21. Kagan BL, Finkelstein A, Colombini M (1981) Diphtheria toxin fragment forms large pores in phospholipid bilayer membranes. Proc Natl Acad Sci USA 78:4950-4954 22. Donovan JJ, Simon MI, Draper RK, Montal M (1981) Diphtheria toxin forms transmembrane channels in planar lipid bilayers. Proc Natl Acad Sci USA 78:172-176 23. Hoch DH, Romero-Mira M, Ehrich BE, Finkelstein A, DasGupta BR, Simpson LL (1985) Channels formed by botulinum, tetanus, and diphtheria toxins in planar lipid bilayers: relevance to translocation of proteins across membranes. Proc Natl Acad Sci USA 82:336-343 24. Papini E, Sandon´a D, Rappuoli R, Montecucco C (1988) On the membrane translocation of diphtheria toxin: at low pH the toxin induces ion channels in cells. EMBO J 7:3353-3359 25. Sandvig K, Olsnes S (1988) Diphtheria toxin-induced channels in Vero cells selective for monovalent cations. J Biol Chem 263:12352-12359 26. O’Keefe DO, Cabiaux V, Choe S, Eisenberg D, Collier RJ (1992) pH-dependent insertion of proteins into membranes: B-chain mutation of diphtheria toxin that inhibits membrane translocation, Glu349 → Lys. Proc Natl Acad Sci USA 89:6202-6206 27. Falnes PO, Madhus IH, Sandvig K, Olsnes S (1992) Replacement of negative by positive charges in the presumed membrane inserted part of diphtheria toxin B fragment: effect on membrane translocation and on formation of channels. J Biol Chem 267:12284-12290 28. vanderSpek J, Cassidy D, Genbauffe F, Huynh P, Murphy JR (1994) An intact transmembrane helix 9 is essential for the efficient delivery of the diphtheria toxin catalytic domain to the cytosol of target cells. J Biol Chem 69:21455-21459 29. Lambotte P, Falmagne P, Capiau C, Zanen J, Ruysschaert J-M, Dirkx J (1980) Primary structure of diphtheria toxin fragment B: structural similarities with lipid-binding domains. J Cell Biol 87:837-840 30. Stenmark H, Ariansen S, Afanasieu BN, Olsnes S (1992) Interactions of diphtheria toxin B fragment with cells. Role of amino and carboxy-terminal regions. J Biol Chem 281:619-625 31. vanderSpek JC, Mindel J, Finkelstein A, Murphy JR (1993) Structure function analysis of the transmembrane domain of the interleukin-2 receptor target fusion toxin DAB389IL-2: The amphipathic helical region of the transmembrane domain is essential for the efficient delivery of the catalytic domain to the cytosol of target cells. J Biol Chem 268:12077-12082 32. Wiedlocha A, Madhus IH, Mach H, Middaugh CR, Olsnes S (1992) Tight folding of acidic fibroblast growth factor prevents its translocation into the cytosol with diphtheria toxin as vector. EMBO J 11:4835-4842 33. Falnes PO, Choe S, Madhus IH, Wilson BA, Olsnes S (1994) Inhibition of membrane translocation of diphtheria toxin
Acknowledgements We thank Larry Cosenza for the ribbon diagram of native diphtheria toxin. JRM and JCvdS are partially supported by grant CA-60934 from the National Cancer Institute. JvdS is a Special Fellow of the Leukemia Society of America.
References 1. Uchida T, Gill DM, Pappenheimer AM, Jr (1971) Mutation in the structural gene for diphtheria toxin carried by temperate phage β. Nature 233:8-11 2. Murphy JR, Pappenheimer AM Jr, Tayart de Borms S (1974) Synthesis of diphtheria tox gene products in Escherichia coli extracts. Proc Natl Acad Sci USA 71:11-15 3. Boquet P, Silverman MS, Pappenheimer AM Jr, Vernon WB (1976) Binding of triton X-100 to diphtheria toxin, crossreacting material 45, and their fragments. Proc Natl Acad Sci USA 73:4449-4453 4. Middlebrook JL, Dorland RB, Leppla SH (1978) Association of diphtheria toxin with Vero cells: demonstration of a receptor. J Biol Chem 253:7325-7330 5. Uchida T, Pappenheimer AM Jr, Greaney R (1973) Diphtheria toxin and related proteins: isolation and properties of mutant proteins serologically related to diphtheria toxin. J Biol Chem 248:3838-3844 6. Hayakawa S, Uchida T, Mekada E, Moynihan MR, Okada Y (1983) Monoclonal antibody against diphtheria toxin: effect on toxin binding and entry into cells. J Biol Chem 258:4311-4317 7. Murphy JR, Bishai W, Borowski M, Miyanohara A, Boyd J, Nagle S (1986) Genetic construction, expression, and melanomaselective cytotoxicity of a diphtheria toxin α-melanocyte stimulating hormone fusion protein. Proc Natl Acad Sci USA 83:8258-8262 8. Greenfield L, Johnson VG, Youle RJ (1987) Mutations in diphtheria toxin separate binding from entry and amplify immunotoxin selectivity. Science 238:536-539 9. Myers DA, Villemez CL (1988) Specific chemical cleavage of diphtheria toxin with hydroxyamine: purification and characterization of the modified proteins. J Biol Chem 263:17122-17127 10. Rolf JM, Eidels L (1993) Structure-function analyses of diphtheria toxin by use of monoclonal antibodies. Infect Immun 61:944-1003 11. Schalken JA, Roebuck AJM, Oomen PPCA, Wagenaar SS, Debruyne FM, Bloemers HPJ, van de Ven WJM (1987) Furin gene expression as a discriminating marker for small cell and non small cell lung carcinomas. J Clin Invest 8:1545-1549 12. Barr PJ (1991) Mammalian subtilisins: the long sought dibasic processing endoproteases. Cell 66:1-3 13. Williams DP, Wen Z, Watson RS, Boyd J, Strom TB, Murphy JR (1990) Cellular processing of the fusion toxin DAB486IL-2 and efficient delivery of diphtheria toxin fragment A to the cytosol of target cells requires Arg194. J Biol Chem 265:20673-20677 14. Klimpel KR, Molloy SS, Thomas G, Leppla S (1992) Anthrax toxin protective antigen is activated by a cell surface protease with the sequence specificity and catalytic properties of furin. Proc Natl Acad Sci USA 89:10277-10281 15. Tsuneoka M, Nakagama K, Hatsuzawa K, Komada M, Kitamura N, Mekada E (1993) Evidence for involvement of furin in cleavage and activation of diphtheria toxin. J Biol Chem 268:26461-26465
266
Targeting diphtheria toxin
34. 35. 36.
37.
38. 39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
A-fragment by internal disulfide bridges. J Biol Chem 269:8402-8407 Choe S, Bennett MJ, Fujii G, Curmi PMG, Kantardjieff KA, Collier RJ, Eisenberg D (1992) The crystal structure of diphtheria toxin. Nature 357:216-222 Moskaug JO, Sandvig K, Olsnes S (1987) Cell-mediated reduction of the interfragment disulfide in nicked diphtheria toxin. J Biol Chem 262:10339-10345 Papini E, Rappuoli R, Muriga M, Montecucco C (1993) Cell penetration of diphtheria toxin. Reduction of the interchain disulfide bridge is the rate-limiting step of translocation into the cytosol. J Biol Chem 268:1567-1574 Mandel R, Ryser H-P, Ghani F, Wu M, Peak D (1993) Inhibition of a reductive function of the plasma membrane by bacitracin and antibodies against protein disulfide isomerase. Proc Natl Acad Sci USA 90:4112-4116 Freedman RB (1989) Protein disulfide isomerase multiple role in the modification of nascent secretory proteins. Cell 57:1069-1072 Bacha P, Murphy JR, Reichlin S (1983) Thyrotropin-releasing hormone-diphtheria toxin-related polypeptide conjugates: potential role of the hydrophobic domain in toxin entry. J Biol Chem 258: 1565-1570 Cawley DB, Herschman HR, Gilliland DG, Collier RJ (1980) Epidermal growth factor-toxin A chain conjugates: EGF-ricin A is a potent toxin while EGF-diphtheria fragment A is nontoxic. Cell 22:563-570 Williams D, Parker K, Bishai W, Borowski M, Genbauffe F, Strom TB, Murphy JR (1987) Diphtheria-toxin receptor binding domain substitution with interleukin-2: genetic construction and properties of a diphtheria toxin-related interleukin-2 fusion protein. Protein Engn 1:493-498 Williams D, Snider CE, Strom TB, Murphy JR (1990) Structure function analysis of IL-2 toxin (DAB486IL-2): fragment B sequences required for the delivery of fragment A to the cytosol of target cells. J Biol Chem 265:11885-11889 Bacha P, Waters C, Williams J, Murphy JR, Strom TB (1988) Interleukin-2 targeted cytotoxicity: selective action of a diphtheria toxin-related interleukin-2 fusion protein. J Exp Med 167:612-622 Waters CA, Schimke P, Snider CE, Itoh K, Smith KA, Nichols JC, Strom TB, Murphy JR (1990) Interleukin-2 receptor targeted cytotoxicity: receptor binding requirements for entry of IL2-toxin into cells. Eur J Immunol 20:785-791 Aullo P, Alcani J, Popoff MR, Klatzman DR, Murphy JR, Boquet P (1992) In vitro effects of a recombinant diphtheria-human CD4 fusion toxin on acute and chronically HIV-1 infected cells. EMBO J 12:921-931 Lakkis F, Steele A, Pacheco-Silva A, Kelley VE, Strom TB, Murphy JR (1991) Interleukin-2 receptor targeted cytotoxicity: genetic construction and properties of diphtheria toxin-related interleukin-4 fusion toxins. Eur J Immunol 21:2253-2258 Wen Z, Tao X, Lakkis F, Kiyokawa T, Murphy JR (1991) Expression, purification, and α-melanocyte stimulating hormone receptor-specific toxicity of DAB-α-MSH fusion toxins. J Biol Chem 266:12289-12293 Shaw JP, Akiyoshi DE, Arrigo DA, Rhoad AE, Sullivan B, Thomas J, Genbauffe FS, Bacha P, Nichols JC (1991) Cytotoxic properties of DAB486EGF and DAB389EGF, epidermal growth factor (EGF) receptor-targeted fusion toxins. J Biol Chem 266:13449-13455
49. Jean L-F, Murphy JR (1992) Diphtheria toxin receptor binding domain substitution with interleukin-6: genetic construction and interleukin-6 receptor specific action of a diphtheria toxinrelated interleukin-6 fusion protein. Protein Engn 4:989-994 50. Sweeney EB, vanderSpek JC, Foss F, Murphy JR (1995) Genetic construction and characterization of DAB389IL-7: A novel agent for the elimination of IL-7 receptor positive cells. (submitted) 51. vanderSpek J, Howland K, Friedman T, Murphy JR (1994) Maintenance of the hydrophobic face of the diphtheria toxin amphipathic transmembrane helix 1 is essential for the efficient delivery of the catalytic domain to the cytosol of target cells. Protein Engn 7:985-989 52. Bennett MJ, Choe S, Eisenberg D (1994) Domain swapping: entangling alliances between proteins. Proc Natl Acad Sci USA 91:3127-3131 53. Waldmann TA (1986) The structure, function, and expression of interleukin-2 receptors on normal and malignant T cells. Science 232:727-732 54. Waldmann TA (1990) The multichain interleukin-2 receptor. A target for immunotherapy in lymphoma, autoimmune disorders, and organ allografts. J Am Med Assoc 263:272-274 55. LeMaistre CF, Meneghetti C, Rosenblum M, Reuben J, Parker K, Shaw J, Woodworth T, Parkinson D (1992) Phase I trial of an interleukin-2 receptor (IL-2R) fusion toxin (DAB486IL-2) in hematologic malignancies expressing the IL-2 receptor. Blood 79:2547-2554 56. Schwartz G, Tepler I, Charette J, Kadin L, Parker K, Woodworth T, Schnipper L (1992) Complete response of a Hodgkin’s lymphoma in a phase I trail of DAB486IL-2. Blood 79:175a 57. Hesketh P, Caguioa P, Koh H, Dewey H, Facada A, McCaffrey R, Parker K, Nylen P, Woodworth T (1993) Clinical activity of a cytotoxic fusion protein in the treatment of cutaneous T cell lymphoma. J Clin Oncol 11:1628-1690 58. LeMaistre CF, Craig FE, Meneghetti C, McMullin B, Parker K, Reuben J, Boldt DH, Rosenblum M, Woodworth T (1993) Phase I trail of a 90-minute infusion of the fusion toxin DAB486IL-2 in hematologic cancers. Cancer Res 53:3930-3934 59. Chatenoud L, Jonker M, Villemain F, Goldstein G, Bach JF (1986) The human immune response to the OKT3 monoclonal antibody is oligoclonal. Science 232:1406-1408 60. Bacha P, Forte SE, Perper SJ, Trentham DE, Nichols JC (1992) Anti-arthritic effects demonstrated by an interleukin-2 receptortargeted cytotoxin (DAB486IL-2) in rat adjuvant arthritis. Eur J Immunol 22:1673-1679 61. Zucker DR, Murphy JR (1984) Monoclonal antibody analysis of diphtheria toxin. I. Localization of epitopes and neutralization of cytotoxicity. Molec Immunol 21:785-793 62. Gottlieb SL, Gilleaudeau P, Johnson R, Estis L, Woodworth T, Gottlieb AB, Krueger JG (1995) Response of psoriasis to a lymphocyte-selective toxin (DAB389IL-2) suggests a primary immune, but not keratinocyte, pathogenic basis. Nature Med 1:442-447 63. Kraulis PJ (1991) MOLESCRIPT: a program to produce both detailed and schematic plots of protein structures. J Appl Crystallogr 24:946-950 64. Kiyokawa T, Williams DP, Snider CE, Strom TB, Murphy JR (1991) Protein engineering of diphtheria toxin-related interleukin-2 fusion toxins to increase biologic potency for high affinity interleukin-2 receptor bearing cells. Protein Engn 4:463-468
267