9 1993 by Humana Press Inc. All rights of any nature whatsoever reserved. 0163-4992/92/21/121-138/$4.60
Rational Immunotherapy with Ribonuclease Chimeras An Approach Toward Humanizing Immunotoxins SUSANNA M. RYBAK,*'I HFNNm R. HOOGFNBOOM,2 D~NNE L. NEWTON,1 JFF C. M. RAUS,2 AND RICHARDJ. YOULE1 ISurgical Neurology Branch, National Institute of Neurological Diseases and Stroke, National Institutes of Health, Bethesda, MD 20892 and 2Dr. L. Willems Institute, 3590 Diepenbeek, Belgium ABSTRACT Members of the pancreatic ribonuclease (RNase) family have diverse activities toward RNA that could cause them to function during host defense and physiological cell death pathways. This activity could be harnessed by coupling RNases to cell binding ligands for the purpose of engineering them into cell-type specific cytotoxins. Therefore, the cytotoxic potential of RNase was explored
*Author to whom all correspondence and reprint requests should be addressed: Laboratory of Biochemical Physiology, National Cancer Institute, National Institutes of Health, Frederick, MD 21702. Cell Biophysics
121
Volume 21, 1992
122
Rybak et al. by linking bovine pancreatic ribonuclease Avia a disulfide bond to human transferrin or antibodies to the transferrin receptor. The RNase hybrid proteins were cytotoxic to K562 human erythroleukemia cells in vitro with an IC50around 10-7M,whereas > 10-4M of native RNase was required to inhibit protein synthesis. Cytotoxicity required both components of the conjugate since excess transferrin or ribonuclease inhibitors added to the medium protected the cells from the transferrin-RNase toxicity.Importantl~ the RNase conjugates were found to have potent antitumor effects in vivo. Chimeric RNase fusion proteins were also developed. F(ab')2-1ike antibody-enzyme fusions were prepared by linking the gene for human RNase to a chimeric antitransferrin receptor heavy chain gene. The antibody enzyme fusion gene was introduced into a transfectoma that secreted the chimeric light chain of the same antibody, and cell lines were cloned that synthesized and secreted the antibody-enzyme fusion protein of the expected size at a concentration of 1-5 ng/mL. Culture supernatants from clones secreting the fusion protein caused inhibition of growth and protein synthesis toward K562 cells that express the human transferrin receptor but not toward a nonhuman derived cell line. Since human ribonucleases coupled to antibodies also exhibited receptor mediated toxicities, a new approach to selective cell killing is provided. This may allow the development of new therapeutics for cancer treatment that exhibit less systemic toxicity and, importantly, less immunogenicity than the currently employed ligand-toxin conjugates. Index Entries: Ribonucleases; immunotoxins; cancer; immunotherapy.
INTRODUCTION
Toxic enzymes from plants and bacteria, such as ricin, diphtheria toxin, and Pseudomonas toxin, have been coupled to antibodies or receptor-binding ligands to generate cell-type-specific killing reagents called immunotoxins (1-3). The injection of immunotoxins containing plant or bacterial proteins into patients was anticipated to elicit an antibody response that would present a major obstacle to the successful application of this technology. Indeed i m m u n e responses against murine monoclonal antibodies (4,5) and antitoxin antibodies have been detected in both animals and humans treated with these reagents (6-8). In that Cel[Biophysics
Volume 21, 1992
Ribonuclease Chimeras
123
regard, efforts to reduce the antigenicity of the antibody portion of immunotoxins are under way. A great deal of work is being directed toward obtaining chimeric (9,10) or humanized antibodies (11,12) for human therapeutics. Recently, results from our laboratory have addressed the immunogenicity of the toxin portion of immunotoxins. Mammalian pancreatic RNase A linked to transferrin (13) or antibodies to the transferrin receptor (14) display receptor-mediated cytotoxicities characteristic of specific-toxin conjugates. The feasibility of using human RNase in the construction of cytotoxic RNase chimeras also has been demonstrated (15). Human RNase in lieu of bacterial or plant toxins could make possible the construction of a new type of immunotherapeutic with significantly less immunogenicity and less systemic toxicity than plant- and bacterial-derived immunotoxins.
RATIONALE FOR PANCREATIC-TYPE RNASES AS TARGETED DRUGS Recent evidence indicates that some extracellular RNases have physiologic roles in specific killing of eukaryotic cells. The style selfincompatibility gene products that prevent fertilization by pollen in the flowering plant Nicotiana alata are RNases (16). Since RNase exerts a physiological role by killing cells in plants, it is intriguing to speculate that RNases may have cytotoxic regulatory roles in other eukaryotic systems. Certainly this may be true for certain species of frog, since an RNase isolated from extracts of Rana pipiens early embryos has antiproliferative/cytotoxic effects toward cancer cells (17,18). The primary sequence of this protein (onconase) is highly homologous to that of the pancreatic RNases (19), and it exhibits RNase activity toward highly polymerized RNA that may be essential for its cytotoxic/antitumor effects (19). Other studies have demonstrated its potential as an anticancer therapeutic in animal models (20). Moreover, phase 1 human clinical trials in patients with pancreatic carcinoma and other solid tumors have recently been completed (personal communication). The physiological roles of extracellular mammalian RNases are unknown. Classically, pancreatic RNase A has been thought to function in digestion of RNA. However, the existence of many members of the RNase A supergene family in a variety of cells and organs has led to the speculation of other functions for RNase, reviewed in (21). In mammals, reports of antitumor effects of certain forms of RNase CelI Biophysics
Volume21, 1992
124
Rybak et al.
A are documented from studies performed in the 1960s and 1970s, reviewed in (22). In addition, bovine seminal RNase, a naturally occurring dimer, exhibits antitumor activity (23,24). As yet there is no direct evidence that human RNases have antitumor effects, but human eosinophils contain large amounts of cytotoxic RNases (25) and are hypothesized to have antineoplastic effects (26). Indeed, the in vivo antitumor effect of interleukin 4 was found to be mediated by a host inflammatory infiltrate composed of eosinophils and macrophages (27). Another recent study suggests that eosinophils are recruited to tumor-developing sites in hamster oral carcinogenesis and predominantly associate with malignant epithelium (28). Eosinophilderived neurotoxin (EDN) and eosinophil-derived cationic protein (ECP), two of the major proteins in the secretory granules of cytotoxic eosinophils, are homologous to pancreatic RNase and possess RNase activity (25). In addition, ECP has pore-forming properties (29), and that has led to the proposal that ECP released from the eosinophil granule inserts itself into the target cell membrane and assists in the transfer of EDN into the cell thereby halting protein synthesis (30). In the sense that the eosinophil may be targeting RNases, our work with RNase conjugates would be mimicking a natural host defense mechanism. Human serum also contains several extracellular RNases (31,32) that are expressed in a tissue-specific manner. Angiogenin, one of these human serum RNases, has been well characterized with regard to several disparate biological activities. It was discovered by fractionating tumor-cell-conditioned medium and following the specific activity to elicit angiogenesis in the chick chorioallentoic membrane assay (33). The structure of angiogenin contains about 65% homology to pancreatic RNase A (34,35) and the active site residues are conserved, yet it has very little activity toward standard substrates for the pancreatic enzyme (36). However, angiogenin is a potent inhibitor of cell-free protein synthesis by ribonucleolytically inactivating some component of the translational machinery in the rabbit reticulocyte lysate (37). Since all of the classical toxins used to construct immunotoxins also affect some cellular component to cause the cessation of protein synthesis and cell death, cell-free protein synthesis inhibitory activity of several members of the pancreatic RNase family was determined (Table 1). The results are interesting in that each of these RNases are potent inhibitors of protein synthesis in the rabbit reticulocyte lysate, even though there is wide variation in the ability of each to hydrolyze typical RNase A substrates enzymatically. Yet, ribonuclease activity is Cell Biophysics
Volume21, 1992
Ribonuclease Chimeras
125
6 2
0
o
+
z
-~
o.i
~
~.~ ~-.i
r~
o
o~
o
o
c~ o
~.~ m ~
~
.~
9 .~.~
i ~~ .~~ -~ Cell B i o p h y s i c s
~
•
o
V o l u m e 21, 1992
126
Rybak et al.
essential for inhibition of cell-free protein synthesis. Therefore, the disparate enzymatic activities toward substrates optimal for bovine pancreatic RNase A suggest that there could be different intracellular RNA targets for these enzymes. In addition, these RNases exhibit different profiles of cellular cytotoxicity. Pancreatic RNase A and angiogenin are present extracellularly in humans, and neither is toxic to cells in culture at concentrations at which EDN and frog onconase have cytotoxic effects. Taken together these results suggest that all of these RNases are capable of cellular cytotoxicity, although not all have the inherent ability to enter the cell to exert the toxic effect. Hence, in addition to pharmacological application, discussed next, these studies may illuminate as yet undiscovered biological activities of some members of this protein super family.
IN VITRO AND IN VlVO EFFECTS OF RNase CHIMERAS Chemical RNase Chimeras RNase conjugates were formed by derivatizing bovine pancreatic RNase A with the heterobifunctional crosslinking reagent N-succinimidyl 3-(2-pyridyldithio)propionate (SPDP) and reacting it with 2-iminothiolane-treated ligand (13,14). The transferrin receptor was targeted initially because it is particularly effective as a ligand for immunotoxins (38,39), since it passages a series of endocytotic compartments functioning to cycle iron-bound transferrin into the cell and apo-transferrin back to the cell surface. Additionally, transferrin receptor-mediated endocytosis is capable of transporting macromolecules into cells (40). RNase A coupled to two different antihuman transferrin receptor monoclonal antibodies (454A12 or 5E-9), or to transferrin itself, acquired cellspecific cytotoxicity directed toward those cell lines expressing the human transferrin receptor (Fig. 1). Moreover, the cytotoxicity of RNase chimeric proteins is not unique to the transferrin receptor. RNase A conjugated to an antibody against the human T-cell-specific antigen CD5 (T101) also was cytotoxic to the appropriate target cell (Fig. 1). Whereas the mechanism of cell killing by mammalian RNase fusion proteins has not been directly demonstrated, degradation of cellular RNA can cause cytotoxicity. A general degradative bacterial RNase Cell Biophysics
Volume 21, 1992
Ribonuclease Chimeras
127
120 o O O
w M Q
100 --
8O
~
\
60 :
r-
40. q
_= Q
[] =
5EgRNase 454RNase
--
Tfn-RNase
9
T101RNase
--
RNase
20, 0
m .mmJ
Io-Y
m m m ~
10-8
m mmm~
m mm.~
m mmm~
10-7 10-1~ 10-5 Concentration (M)
m ..m~
10-4
10-3
Fig. 1. Inhibition of prote~n synthesis in cultured cells by l ~ a s e chimeras vs [~Tase. Protein synthesis was detem-~_ned as described in (13).
RNase A chimeras chemically constructed with transferrin (Tfn), antibodies to the human transferrin receptor (5E9, 454), or to the human T-cell antigen (T101) are identified to the right of the figure. gene genetically fused to the gene for Pseudomonas toxin inhibited protein synthesis in cells that were refractory to the toxin partly by degrading R N A (41). Furthermore, on microinjection into oocytes, RNase A is as potent as ricin or diphtheria toxin in causing inhibition of protein synthesis, and the inhibition correlates directly to degradation of cellular RNA (42). RNase A (Fig. 1) or RNase A mixed with transferrin (13) is virtually nontoxic to mammalian cells in culture inhibiting protein synthesis only at millimolar concentrations. This inhibition appears to be reversible, since RNase A does not kill cells when assessed in clonogenic assays that measure cell killing directly (13). Therefore, RNase A must be linked to a ligand to acquire cytotoxicity. Furthermore, both component proteins of the chimera are required for toxicity (Table 2). Excess transferrin interferes with toxicity whether measured by inhibition of cellular protein synthesis or surviving clonogenic units, implying that toxicity is receptor-mediated. Inhibitors of RNase also interfere with the toxicity of the conjugate, thus addressing the role of chemically linked RNase A (Table 2). In view of these results, it is presumed that chimeric RNase enters an intracellular compartment by a receptor-mediated event after which the entire conjugate and/or the RNase alone transits to the cytosol to kill the cell. Cell Biophysics
Volume 21, 1992
128
Rybak et al. Table 2 Specific Reversal of Chimeric RNase Activity Surviving clonogenic units % of Control Tfn-RNase, mM 0 1.4
-Tfn
+Tfn
100 0.002
62 31
Protein synthesis Tfn-RNase
-PRI
+PRI
20
36
100
200
3
100
aAssay protocols described in and data adapted from ref. 13. PRI and Tfn stand for placental RNase inhibitor and transferrin, respectively.
The kinetics of cellular toxicity of plant or bacterial toxins exhibit an initial lag period in which no protein synthesis inhibition is observed (reviewed in ref. 43). Both the length of the lag and the rate of decrease in protein synthesis are concentration-dependent. Similarly, conjugates made with RNase A also exhibit dose-dependent lag times followed by log linear decreases in protein synthesis (13). The rate-limiting step in cytotoxicity of classical immunotoxins is considered to be the translocation of the toxin to the cytosol. Since RNase A directly microinjected into oocytes is roughly 25,000x more potent in abolishing oocyte protein synthesis than RNase A conjugates constructed with chemical linkers, one model would predict that efficient entry of RNase A into the appropriate cellular compartment is hampered. Although bovine pancreatic RNase A is not cytotoxic, a homolog isolated from frogs is a natural toxin (17). Onconase, discussed in the preceeding section, has cytotoxic (51,52) and antitumor properties (19) that appear to be dependent on its RNase activity. Thus it appears that onconase is capable of entering the cytosol. Interestingly and importantly, the ability to enter cells does not result in a more potent conjugate when onconase Cell Biophysics
Volurne 21, 1992
R ibonuclease Chimeras
129
is chemically linked to transferrin or human antitransferrin receptor antibodies (53). These results suggest that optimizing the coupling reaction or the use of different chemical linkers will be necessary to construct more potent chemically linked RNase chimeras.
Recombinant Human RNase Chimeras The lack of sufficient protein has precluded the chemical conjugation of human RNases. Thus it was important to determine whether RNases could be expressed in a functional form as chimeric fusion proteins and whether they would also acquire directed cytotoxic properties when fused to an antibody. The human RNases depicted in Table 1 are both possible candidates for the construction of human RNase chimeras. Based on our studies with nontoxic bovine RNase A compared to the toxic frog enzyme in the construction of RNase chimeras, one might predict that nontoxic human RNases, such as angiogenin, may make fusion proteins with the best therapeutic index. Thus, initial studies were performed with a recombinant form of angiogenin (42) referred to as ANG. The chimeric mouse-human antibody to the transferrin receptor that had been well characterized for its ability to target tumor necrosis factor (44,45) was used to construct ANG fusion proteins. The gene for ANG was modified and cassetted into an expression vector containing the chimeric antibody gene at the 5' region of the CH2 domain of the antibody, thus leaving the hinge region unaffected and dimerization of the heavy chain possible (15). The junctions between the antibody heavy chain and the ANG gene are depicted at the top of Fig. 2. An antitransferrin receptor chimeric lightchain-producing cell line E12B5 (46) was transfected with the vector containing the CH2ANG gene. After selection for the presence of the gpt gene, culture supernatants of clones testing positive for human IgG were followed for reproducible human IgG activity. Of these, clone CH2ANG was selected for further characterization. The amount of secreted CH2ANG ranged from I to 5 ng/mL as determined in a human IgG-detecting ELISA and from I to 2 ng/mL when ANG was detected by ELBA. Unfortunately, the level of secretion of this chimeric antibody linked to ANG or TNF was far lower than that described for other chimeric enzymes (47,48). It is probable that the antibody secretory pathway used by these constructs did not protect the cell from cytotoxicity to the same extent as would the normal bioCell Biophysics
Volume 21, 1992
130
Rybak et al.
Chimeric CH2Ang CK
VL
_l
I VH
CHI
I
H CH2
ANG
I I illlllllllllllllllllllllllllllllllllllllll
I
f
--a-~ntibody ..... CH2- junction ............ ANG........ A
P
GCA CCT
E GAA
F TTC
Q D N S CAG GAT AAC TCC
--EcoRl--
Fig. 2. Sequences around the junction between antibody heavy chain in the CH2 domain and the 5' end of the sequence coding for mature ANG in the CH2ANG construct. E and F are amino acids introduced by PCR modification of the ANG gene prior to cloning. Single-letter amino acid code is used. VH, mouse variable regions of E6 antibody gene; CH1 and CH2, constant regions I and 2, respectively, of the heavy-chain gene; H, hinge region of heavy-chain gene; ANG, DNA coding for human angiogenin is shown in a shaded box. synthetic pathway for these proteins and high-level expression possibly caused the self-selection of low producers. The extremely low yield of protein precluded the purification of the ANG chimera to homogeneity. However, partially purified material was found to express characteristic in vitro activities of the native RNase (Fig. 3A). CH2ANG inhibited protein synthesis in the rabbit reticulocyte lysate, and this was reversed by including an inhibitor of ANG. Importantly, medium containing the fusion protein killed human leukemia cells that express the transferrin receptor (Fig. 3B), while sparing cells that did not Cell Biophysics
Volume 21, 1992
Ribonuclease Chimeras A
131
120
o
100
~
6O
0
r-
~ if)
411
"3
20
o.
0
Control
B
CH2ANG 40 ng
CH2ANG 80 ng
CH2ANG 80 ng+PRI
10001 100
0
5
10
1"5
DAYS IN C U L T U R E
Fig. 3. A: Effect of CH2ANG on protein synthesis in the rabbit reticulocyte lysate in the absence or presence of ANG inhibitor (PRI). B: Growth of K562 cells in transfectoma culture supematant. K562 cells were grown in I mL of growth medium (--B--) or I mL of growth medium that contained 1-2 ng/mL of CH2ANG (me--). The data are adapted from ref. 15. Cell Biophysics
Volume 21, 1992
132
Rybak et al.
recognize the human antibody (15). In addition, cytotoxicity was completely blocked by the daily addition of excess antibody (15). Thus, human RNase can be expressed as a recombinant fusion protein that not only retains some of its individual biological activities, but also acquires cytotoxicity similar to the chemically linked bovine RNase A. Additionally, the construction of a unique linkage between RNase and a cell-binding ligand eliminates the heterogeneity of chemically linked antibody-RNase conjugates. This may have contributed to the increased potency of the recombinant antibody fusion protein. Based on immunoreactivity, the recombinant RNase chimera was about 1000 times more potent than chemically linked RNase hybrids. The established host defense activities of EDN and ECP make them intriguing candidates for immunotherap~ since the concept of targeting natural biological response modifiers is appealing. In this regard, an antibody-tumor necrosis factor hybrid molecule has been expressed and secreted by myeloma cells (44,45), and the TNF fusion protein was found to be more toxic to specific target cells than the native cytokine. For these reasons, studies are ongoing to express RNase as single-chain antibody fusion proteins in E. coli. It is hoped that this methodology will allow the production of sufficient chimeric protein to establish clinical relevance.
In Vivo Antitumor Effects of RNase Chimeras Antitransferrin receptor RNase A conjugates were evaluated for antitumor effects in vivo (14). U251 human glioma tumors were grown in the flanks of nude mice until a tumor size of 0.5-1.0 cm diameter was reached. The tumors were injected intratumorally with four doses of saline, RNase A conjugate, or a nonconjugated mixture of RNase A and the antibody (Fig. 4). Only the RNase A chimera inhibited tumor growth, and the inhibitor effect was maintained up to 30 d when the experiment was terminated. The in vivo effect of the conjugate mimicks in vitro activity in that the unconjugated mixture had no effect on tumor growth. These in vivo data confirm that RNase must be linked to the antibody to acquire cytotoxicity. Although chemically linked RNase chimeras are considerably less active to inhibit in vitro cellular protein synthesis than hybrid proteins constructed with antibodies and classical toxins (13,14), they are sur-
Cell Biophysics
Volume 21, 1992
Ribonuclease Chimeras
133
A
I,LI
1.50
+1 CO
I5 o
1.00 I= "6
E =
l--
0.50
C G)
0.00
0
10
20
30
Days
Fig. 4. Tumoricidal activity of an anitibody-RNase chimera directed against the human transferrin receptor. Solid U251 human glioblastoma tumors were established in the flanks of nude mice. Animals were treated with direct intratumoral injections every other day for a total of four doses. Saline (4, N = 10); RNase mixed with 5E9 antibody (44 and 210 pmol, respectively; m, N = 10), and 5E9 RNase chimera (27-29 pmol; [3, N = 10). Reproduced with permission from (14). prisingly potent in vivo. The antitumor effect of the antitransferrin receptor antibody-RNase A conjugate was equal to that of an antitransferrin receptor ricin A-chain immunotoxin (454A12-rRA) that has been used in a human clinical trial (Laske et al., manuscript in preparation). The disparity in the in vitro and in vivo potency of chemically linked RNase A chimeras hints at factors that may amplify the effect of RNase containing drugs in vivo. If so, RNase chimeras may help to compensate for such problems as antigen heterogeneity, which limits access of immunotherapeutics directed to internalizing antigens.
Cell Biophysics
Volume21, 1992
134
Rybak et al.
SUMMARY Targeting drugs for cancer therapy with antibodies directed to tumor cells offers the hope of selective toxic effects. Targeted therapy with drugs constructed with toxic proteins conjugated to antibodies might offer the best possibility for success since (1) they act catalytically rather than stoichiometrically, and (2) the requirement for cellular internalization of the toxin to kill the target cell helps to assure that nearby normal cells are not killed as they would be with antibodies conjugated to radioisotopes or surface-acting cytolytic agents. In vivo therapy with immunotoxins has been disappointing in the initial trials, but these trials were designed to test treatment-related toxicities and not efficacy. In this regard, many facets of immunotoxin therapy have been discovered that were not predicted from animal models, and experimental protocols to address these problems are being designed. In addition, future advances in immunology coupled to advances in constructing recombinant immunotoxins will contribute to the increasing success of this therapy as a standard treatment of cancer alone or combined with other therapeutic modalities. To date, only nonmammalian toxins have been used to construct immunotoxins, and human therapy with these reagents has been limited by problems of immunogenicity and clinical trial suspending toxicities (6,49,50). Adopting a rational approach toward immunotherapy by using human RNase could minimize two of the major problems identified in clinical trials, thus contributing to the successful application of this form of directed drug therapy.
REFERENCES 1. Gilliland, D. G., Steplewski, Z., Collier, R. J., Mitchell, K, Chang, T., and Koprowski, H. (1980) Antibody directed cytotoxic agents: use of monoclonal antibody to direct the action of toxin A chains to colorectal carcinoma cells. Proc. Natl. Acad. Sci. USA 77, 4539-4543. 2 Krolick, K. A., ViUemez, C., Isakson, P., Urh, J., and Vitetta, E. (1980) Selective killing of neoplastic B cells by antibodies coupled to the A chain of ricin. Proc. Natl. Acad. Sci. USA 77, 5419-5423. 3. Youle, R. J. and Neville, D. M. (1980) Anti-Thy 1.2 monoclonal antibody linked to ricin is a potent cell-type-specific toxin. Proc. Natl. Acad. Sci. USA 77, 5483-5486. Cell Biophysics
Volume 21, 1992
Ribonuclease Chimeras
135
4. Sawler, D. L., Bartholomew, R. M., Smith, L. M., and Dillman, R. (1985) Human immune response to multiple injections of murine monoclonal IgG. J. Immunol. 135,1530-1535. 5. Schroff, R. W., Foon, K. A., Beatty, S. M., Oldham, R., and Morgan, A. (1985) Human anfi-murine immunoglobulin response in patients receiving monoclonal antibody therapy. Cancer Res. 45, 879-885. 6. Rybak, S. M. and Youle, R. J. (1991) Clinical use of immunotoxins: monoclonal antibodies conjugated to protein toxins. Immunol. Allergy Clin. North Am. 11:2, 359-380. 7. Harkonen, S., Stoudemire, J., Mischak, R., Spetler, L., Lopez, H., and Scannon, P. (1987) Toxicity and immunogenicity of monoclonal antimelanoma antibody-ricin A chain immunotoxins in rats. Cancer Res. 47, 1377-1385. 8. Hertler, A. (1988) Human immune response to immunotoxins, in Immunotoxins (Frankel, A. E., ed.), Kluwer Academic, Boston/Dordrecht/ Lancaster, pp. 475-480. 9. Boulianne, G. L., Houzumi, N., and Schulman, M. J. (1985) Production of functional chimeric mouse/human antibodies. Nature (Lond.) 643-646. 10. Morrison, S. L., Johnson, M. J., Herzenberg, L. A., and Oi, V. T. (1984) Chimeric human antibody molecules: mouse antigen-binding domains with human constant regions. Proc. Natl. Acad. Sci. USA 81, 6851-6855. 11. Jones, P. T., Dear, P. H., Foote, J., Neuberger, M. S., and Winter, G. (1986) Replacing the complementary-determining regions in a human antibody with those from a mouse. Nature (London) 321, 522-525. 12. Riechmann, L., Clark, M., Waldmann, H., and Winter, G. (1988) Reshaping human antibodies for therapy. Nature (Lond.) 332, 323-327. 13. Rybak, S. M., Saxena, S. K., Ackerman, E. J., and Youle, R. J. (1991) Cytotoxic potential of RNase and RNase hybrid proteins. J. Biol. Chem. 266, 21,202-21,207. 14. Newton, D. L., Ilercil, O., Laske, D. W., Oldfield, E., Ryback, S. M., and Youle, R. J. (1992) Cytotoxic ribonuclease chimeras: targeted tumoricidal activity in vitro and in vivo. J. Biol. Chem., submitted. 15. Rybak, S. M., Hoogenboom, H. R., Meade, H., Ravs, J. C., Schwartz, D., and Youle, R. J. (1992) Humanization of immunotoxins. Proc. Natl. Acad. Sci. USA 89, 3165-3169. 16. McClure, B. A., Haring, V., Ebert, P. R., Anderson, M. A., Simpson, R. J., Sakiyama, F., and Clarke, A. E. (1989) Style self-incompatibility gene products of Nicotiana alata are ribonucleases. Nature (Lond.) 342, 955-957. Cell Biophysics
Volume 21, 1992
136
R y b a k et al.
17. Darzynkiewicz, Z., Carter, S. P., Mikulski, S. M., Ardelt, W., and Shogen, K. (1988) Cytostatic and cytotoxic effects of Pannon (P-30 protein) a novel anti-cancer agent. Cell Tissue Kinetics 21, 169-182. 18. Mikulski, S. M., Viera, A., Ardelt, W., Menduke, H., and Shogen, K. (1990) Tamoxifen and trifluoroperazine (stelazine) potentiate cytostatic/cytotoxic effects of P-30 protein, a novel protein possessing anti-tumor activity. Cell Tissue Kinetics 23, 237-246. 19. Ardelt, A., Mikulski, S. M., and Shogen, K. (1991) Amino acid sequence of an anti-tumor protein from Rana pipiens oocytes and early embryos. J. Biol. Chem. 256, 245-251. 20. Mikulski, S. M., Ardelt, W., Shogen, K., Vernstein, E. H., and Menduke, H. (1990) Striking increase of survival of mice bearing M109 Madison Carcinoma treated with a novel protein from amphibian embryos. J. Natl. Cancer Inst. 82, 151-153. 21. Beintema, J., Schuller, C., Irie, M., and Carsana, A. (1988) Molecular evolution of the ribonuclease superfamily. Prog. Biophys. Mol. Biol. 51, 165-192. 22. Roth, J. S. (1963) Ribonuclease activity and cancer: a review. Cancer Res. 23, 657-666. 23. Matousek, J. (1973) The effect of bovine seminal ribonuclease (AS RNase) on cells of Crocker tumour in mice. Experientia 29, 858,859. 24. Vescia, S., Tramontano, D., Augusti-Tocco, G., and D'Allessio, G. (1980) In vitro studies on selective inhibition of tumor cell growth by seminal ribonuclease. Cancer Res. 40, 3740-3744. 25. Slifman, N. R., Loegering, D. A., McKean, D. J., and Gleich, G. J. (1986) Ribonuclease activity associated with human eosinophil-derived neurotoxin and eosinophil cationic protein. J. Immunol. 137, 2913-2917. 26. Spry, C. (1988) Eosinophils: A Comprehensive Review and Guide to the Scientific and Medical Literature. Oxford University Press, Oxford, New York. 27. Tepper, R. I., Pattengale, P. K., and Leder, P. (1989) Murine interleukin4 displays potent anti-tumor activity in vivo. Cell 57, 503-512. 28. Ghiabi, M., Gallagher, G. T., and Wong, D. T. (1992) Eosinophils, tissue eosinophilia and eosinophil-derived transforming growth factor A in hamster oral carcinogenesis. Cancer Res. 52, 389-893. 29. Young, J., Peterson, C., Venge, P., and Cohn, Z. A. (1986) Mechanism of membrane damage mediated by human eosinophil cationic protein. Nature (Lond.) 321, 613-616. 30. Rosenberg, H. F., Tenen, D. G., and Ackerman, S. J. (1989) Molecular cloning of the human eosinophil-derived neurotoxin: a member of the ribonuclease gene family. Proc. Natl. Acad. Sci. USA 86, 4460--4464. Cell Biophysics
Volume 21, 1992
Ribonuclease Chimeras
137
31. Reddi, E. K. (1975) Nature and possible significance of human serum ribonuclease. Biochem. Biophys. Res. Commun. 67, 110-118. 32. Blank, A., Dekker, C., Schieven, G., Sugiyama, R., and Thelen, M. (1981) Human Body Fluid Ribonucleases: Detection, Interrelationships and Significance. IRL, London, pp. 203-209. 33. Fett, J. W., Strydom, D. J., Lobb, R. R., Alderman, E. M., Bethune, J. L Riordan, J. F., and Vallee, B. L. (1985) Isolation and characterization of angiogenin an angiogenic protein from human carcinoma cells. Biochemistry 24, 5480-5485. 34. Strydom, D. J., Fett, J. W., Lobb, R. R., Alderman, E. M., Bethune, J. L Riordan, J. F., and Vallee, B. L. (1985) Amino acid sequence of human tumor derived angiogenin. Biochemistry 24, 5486-5494. 35. Kurachi, K., Davie, E. W., Strydom, D. J., Riordan, J. F., and Vallee, B. L. (1985) Sequence of the cDNA and gene for angiogenin, a human angiogenesis factor. Biochemistry 24, 5494-5499. 36. Shapiro, R., Riordan, J. F., and Vallee, B. L. (1986) Characteristic ribonucleolytic activity of human angiogenin. Biochemistry 25, 3527-3531. 37. St. Clair, D. K., Rybak, S. M., Riordan, J. F., and Vallee, B. L.(1987) Angiogenin abolishes cell-free protein synthesis by specific ribonucleolytic inactivation of ribosomes. Proc. Natl. Acad. Sci. USA 84, 8330-8334. 38. O'Keefe, D. O. and Draper, R. K. (1986) Characterization of a transfertin-diphtheria toxin conjugate. J. Biol. Chem. 260, 932-937. 39 Raso, V. and Basala, M. (1984) A highly cytotoxic human transferrinricin A chain conjugate used to select receptor-modified cells. J. Biol. Chem. 259, 1143-1149. 40. Wagner, E., Zenke, M., Cotten, M., et al. (1990) Transferrin-polycation conjugates as carriers for DNA uptake into cells. Proc. Natl. Acad. Sci. USA 87, 3410-3414. 41. Prior, T. I., Fitzgerald, D. J., and Pastan, I. (1991) Bamase Toxin: a new chimeric toxin composed of pseudomonas exotoxin A and Barnase. Cell 64, 1017-1023. 42. Saxena, S. K., Rybak, S. M., Winkler, G., Meade, H. M., Mc Gray, P., Youle, R. J., and Ackerman, E. J. (1991) Comparison of RNases and toxins on injection into Xenopus oocytes. J. Biol. Chem. 266, 21,20821,214. 43. Hudson, T. H. and Neville, D. M. (1988) Enhancement of immunotoxin action: Manipulation of the cellular routing of proteins, in Immunotoxins (Frankel, A. E., ed.), Kluwar Academic, Boston/Dordrecht/Lancaster, pp. 371-389. Celt Biophysics
Voturne 21, 1992
138
R y b a k et al.
44. Hoogenboom, H. R., Raus, J. M., and Volckaert, G. (1991) Targeting of tumor necrosis factor to tumor cells: secretion by myeloma cells of a genetically engineered antibody-tumor necrosis factor hybrid molecule. Biochem. Biophys. Acta 1096, 345-354. 45. Hoogenboom, H. R., Volckaert, G., and Raus, J. M. (1991) Construction and expression of antibody-tumor necrosis factor fusion proteins. Mol. Immunol. 28, 1027-1037. 46. Hoogenboorn, H. R., Raus, J. M., and Volckaert, G. (1990) Cloning and expression of a chimeric antibody directed against the human transferrin receptor. J. Immunol. 144, 3211-3217. 47. Casadei, J., Powell, M. J., and Kenten, J. H. (1990) Expression and secretion of aequorin as a chimeric antibody by means of a mammalian expression vector. Proc. Natl. Acad. Sci. USA 87, 2047-2051. 48. Williams, G. T. and Neuberger, M. S. (1986) Production of antibody tagged enzymes by myeloma cells: application to DNA polymerase I Klenow fragment. Gene 43, 319-324. 49. Pastan, I., Willingham, M., and FitzGerald, D. (1986) Immunotoxins. Cell 47, 641-648. 50. Vitetta, E. S., Fulton, R. J., and May, R. D. (1987) Redesigning nature's poisons to create anti-tumor reagents. Science 238, 1098-1104. 51. Newton, D. L., Walbridge, S., Mikulski, S. M., Ardelt, K. S., Ackerman, S. J., Rybak, S. M., and Youle, R. J. (1983) Toxicity of an antitumor ribonuclease to Purkinje neurons. J. Neuroscience, in press. 52. Wu, Y. N., Mikulski, S. M., Ardelt, W., Rybak, S. R., and Youle, R. J. (1993) A cytotoxic ribonuclease. J. Biol. Chem. 268, 10,686-10,693. 53. Rybak, S. M., Newton, D. L., Mikulski, S. M., Viera, A., and Youle, R. J. (1993) Cytotoxic onconase and ribonuclease A chimeras. Drug Targeting and Delivery, in press.
Cell Biophysics
Volume 21, 1992
Rational Immunotherapy with Ribonuclease Chimeras An Approach Toward Humanizing Immunotoxins SUSANNA M. RYBAK,*'I HFNNm R. HOOGFNBOOM,2 D~NNE L. NEWTON,1 JFF C. M. RAUS,2 AND RICHARDJ. YOULE1 ISurgical Neurology Branch, National Institute of Neurological Diseases and Stroke, National Institutes of Health, Bethesda, MD 20892 and 2Dr. L. Willems Institute, 3590 Diepenbeek, Belgium ABSTRACT Members of the pancreatic ribonuclease (RNase) family have diverse activities toward RNA that could cause them to function during host defense and physiological cell death pathways. This activity could be harnessed by coupling RNases to cell binding ligands for the purpose of engineering them into cell-type specific cytotoxins. Therefore, the cytotoxic potential of RNase was explored
*Author to whom all correspondence and reprint requests should be addressed: Laboratory of Biochemical Physiology, National Cancer Institute, National Institutes of Health, Frederick, MD 21702. Cell Biophysics
121
Volume 21, 1992
122
Rybak et al. by linking bovine pancreatic ribonuclease Avia a disulfide bond to human transferrin or antibodies to the transferrin receptor. The RNase hybrid proteins were cytotoxic to K562 human erythroleukemia cells in vitro with an IC50around 10-7M,whereas > 10-4M of native RNase was required to inhibit protein synthesis. Cytotoxicity required both components of the conjugate since excess transferrin or ribonuclease inhibitors added to the medium protected the cells from the transferrin-RNase toxicity.Importantl~ the RNase conjugates were found to have potent antitumor effects in vivo. Chimeric RNase fusion proteins were also developed. F(ab')2-1ike antibody-enzyme fusions were prepared by linking the gene for human RNase to a chimeric antitransferrin receptor heavy chain gene. The antibody enzyme fusion gene was introduced into a transfectoma that secreted the chimeric light chain of the same antibody, and cell lines were cloned that synthesized and secreted the antibody-enzyme fusion protein of the expected size at a concentration of 1-5 ng/mL. Culture supernatants from clones secreting the fusion protein caused inhibition of growth and protein synthesis toward K562 cells that express the human transferrin receptor but not toward a nonhuman derived cell line. Since human ribonucleases coupled to antibodies also exhibited receptor mediated toxicities, a new approach to selective cell killing is provided. This may allow the development of new therapeutics for cancer treatment that exhibit less systemic toxicity and, importantly, less immunogenicity than the currently employed ligand-toxin conjugates. Index Entries: Ribonucleases; immunotoxins; cancer; immunotherapy.
INTRODUCTION
Toxic enzymes from plants and bacteria, such as ricin, diphtheria toxin, and Pseudomonas toxin, have been coupled to antibodies or receptor-binding ligands to generate cell-type-specific killing reagents called immunotoxins (1-3). The injection of immunotoxins containing plant or bacterial proteins into patients was anticipated to elicit an antibody response that would present a major obstacle to the successful application of this technology. Indeed i m m u n e responses against murine monoclonal antibodies (4,5) and antitoxin antibodies have been detected in both animals and humans treated with these reagents (6-8). In that Cel[Biophysics
Volume 21, 1992
Ribonuclease Chimeras
123
regard, efforts to reduce the antigenicity of the antibody portion of immunotoxins are under way. A great deal of work is being directed toward obtaining chimeric (9,10) or humanized antibodies (11,12) for human therapeutics. Recently, results from our laboratory have addressed the immunogenicity of the toxin portion of immunotoxins. Mammalian pancreatic RNase A linked to transferrin (13) or antibodies to the transferrin receptor (14) display receptor-mediated cytotoxicities characteristic of specific-toxin conjugates. The feasibility of using human RNase in the construction of cytotoxic RNase chimeras also has been demonstrated (15). Human RNase in lieu of bacterial or plant toxins could make possible the construction of a new type of immunotherapeutic with significantly less immunogenicity and less systemic toxicity than plant- and bacterial-derived immunotoxins.
RATIONALE FOR PANCREATIC-TYPE RNASES AS TARGETED DRUGS Recent evidence indicates that some extracellular RNases have physiologic roles in specific killing of eukaryotic cells. The style selfincompatibility gene products that prevent fertilization by pollen in the flowering plant Nicotiana alata are RNases (16). Since RNase exerts a physiological role by killing cells in plants, it is intriguing to speculate that RNases may have cytotoxic regulatory roles in other eukaryotic systems. Certainly this may be true for certain species of frog, since an RNase isolated from extracts of Rana pipiens early embryos has antiproliferative/cytotoxic effects toward cancer cells (17,18). The primary sequence of this protein (onconase) is highly homologous to that of the pancreatic RNases (19), and it exhibits RNase activity toward highly polymerized RNA that may be essential for its cytotoxic/antitumor effects (19). Other studies have demonstrated its potential as an anticancer therapeutic in animal models (20). Moreover, phase 1 human clinical trials in patients with pancreatic carcinoma and other solid tumors have recently been completed (personal communication). The physiological roles of extracellular mammalian RNases are unknown. Classically, pancreatic RNase A has been thought to function in digestion of RNA. However, the existence of many members of the RNase A supergene family in a variety of cells and organs has led to the speculation of other functions for RNase, reviewed in (21). In mammals, reports of antitumor effects of certain forms of RNase CelI Biophysics
Volume21, 1992
124
Rybak et al.
A are documented from studies performed in the 1960s and 1970s, reviewed in (22). In addition, bovine seminal RNase, a naturally occurring dimer, exhibits antitumor activity (23,24). As yet there is no direct evidence that human RNases have antitumor effects, but human eosinophils contain large amounts of cytotoxic RNases (25) and are hypothesized to have antineoplastic effects (26). Indeed, the in vivo antitumor effect of interleukin 4 was found to be mediated by a host inflammatory infiltrate composed of eosinophils and macrophages (27). Another recent study suggests that eosinophils are recruited to tumor-developing sites in hamster oral carcinogenesis and predominantly associate with malignant epithelium (28). Eosinophilderived neurotoxin (EDN) and eosinophil-derived cationic protein (ECP), two of the major proteins in the secretory granules of cytotoxic eosinophils, are homologous to pancreatic RNase and possess RNase activity (25). In addition, ECP has pore-forming properties (29), and that has led to the proposal that ECP released from the eosinophil granule inserts itself into the target cell membrane and assists in the transfer of EDN into the cell thereby halting protein synthesis (30). In the sense that the eosinophil may be targeting RNases, our work with RNase conjugates would be mimicking a natural host defense mechanism. Human serum also contains several extracellular RNases (31,32) that are expressed in a tissue-specific manner. Angiogenin, one of these human serum RNases, has been well characterized with regard to several disparate biological activities. It was discovered by fractionating tumor-cell-conditioned medium and following the specific activity to elicit angiogenesis in the chick chorioallentoic membrane assay (33). The structure of angiogenin contains about 65% homology to pancreatic RNase A (34,35) and the active site residues are conserved, yet it has very little activity toward standard substrates for the pancreatic enzyme (36). However, angiogenin is a potent inhibitor of cell-free protein synthesis by ribonucleolytically inactivating some component of the translational machinery in the rabbit reticulocyte lysate (37). Since all of the classical toxins used to construct immunotoxins also affect some cellular component to cause the cessation of protein synthesis and cell death, cell-free protein synthesis inhibitory activity of several members of the pancreatic RNase family was determined (Table 1). The results are interesting in that each of these RNases are potent inhibitors of protein synthesis in the rabbit reticulocyte lysate, even though there is wide variation in the ability of each to hydrolyze typical RNase A substrates enzymatically. Yet, ribonuclease activity is Cell Biophysics
Volume21, 1992
Ribonuclease Chimeras
125
6 2
0
o
+
z
-~
o.i
~
~.~ ~-.i
r~
o
o~
o
o
c~ o
~.~ m ~
~
.~
9 .~.~
i ~~ .~~ -~ Cell B i o p h y s i c s
~
•
o
V o l u m e 21, 1992
126
Rybak et al.
essential for inhibition of cell-free protein synthesis. Therefore, the disparate enzymatic activities toward substrates optimal for bovine pancreatic RNase A suggest that there could be different intracellular RNA targets for these enzymes. In addition, these RNases exhibit different profiles of cellular cytotoxicity. Pancreatic RNase A and angiogenin are present extracellularly in humans, and neither is toxic to cells in culture at concentrations at which EDN and frog onconase have cytotoxic effects. Taken together these results suggest that all of these RNases are capable of cellular cytotoxicity, although not all have the inherent ability to enter the cell to exert the toxic effect. Hence, in addition to pharmacological application, discussed next, these studies may illuminate as yet undiscovered biological activities of some members of this protein super family.
IN VITRO AND IN VlVO EFFECTS OF RNase CHIMERAS Chemical RNase Chimeras RNase conjugates were formed by derivatizing bovine pancreatic RNase A with the heterobifunctional crosslinking reagent N-succinimidyl 3-(2-pyridyldithio)propionate (SPDP) and reacting it with 2-iminothiolane-treated ligand (13,14). The transferrin receptor was targeted initially because it is particularly effective as a ligand for immunotoxins (38,39), since it passages a series of endocytotic compartments functioning to cycle iron-bound transferrin into the cell and apo-transferrin back to the cell surface. Additionally, transferrin receptor-mediated endocytosis is capable of transporting macromolecules into cells (40). RNase A coupled to two different antihuman transferrin receptor monoclonal antibodies (454A12 or 5E-9), or to transferrin itself, acquired cellspecific cytotoxicity directed toward those cell lines expressing the human transferrin receptor (Fig. 1). Moreover, the cytotoxicity of RNase chimeric proteins is not unique to the transferrin receptor. RNase A conjugated to an antibody against the human T-cell-specific antigen CD5 (T101) also was cytotoxic to the appropriate target cell (Fig. 1). Whereas the mechanism of cell killing by mammalian RNase fusion proteins has not been directly demonstrated, degradation of cellular RNA can cause cytotoxicity. A general degradative bacterial RNase Cell Biophysics
Volume 21, 1992
Ribonuclease Chimeras
127
120 o O O
w M Q
100 --
8O
~
\
60 :
r-
40. q
_= Q
[] =
5EgRNase 454RNase
--
Tfn-RNase
9
T101RNase
--
RNase
20, 0
m .mmJ
Io-Y
m m m ~
10-8
m mmm~
m mm.~
m mmm~
10-7 10-1~ 10-5 Concentration (M)
m ..m~
10-4
10-3
Fig. 1. Inhibition of prote~n synthesis in cultured cells by l ~ a s e chimeras vs [~Tase. Protein synthesis was detem-~_ned as described in (13).
RNase A chimeras chemically constructed with transferrin (Tfn), antibodies to the human transferrin receptor (5E9, 454), or to the human T-cell antigen (T101) are identified to the right of the figure. gene genetically fused to the gene for Pseudomonas toxin inhibited protein synthesis in cells that were refractory to the toxin partly by degrading R N A (41). Furthermore, on microinjection into oocytes, RNase A is as potent as ricin or diphtheria toxin in causing inhibition of protein synthesis, and the inhibition correlates directly to degradation of cellular RNA (42). RNase A (Fig. 1) or RNase A mixed with transferrin (13) is virtually nontoxic to mammalian cells in culture inhibiting protein synthesis only at millimolar concentrations. This inhibition appears to be reversible, since RNase A does not kill cells when assessed in clonogenic assays that measure cell killing directly (13). Therefore, RNase A must be linked to a ligand to acquire cytotoxicity. Furthermore, both component proteins of the chimera are required for toxicity (Table 2). Excess transferrin interferes with toxicity whether measured by inhibition of cellular protein synthesis or surviving clonogenic units, implying that toxicity is receptor-mediated. Inhibitors of RNase also interfere with the toxicity of the conjugate, thus addressing the role of chemically linked RNase A (Table 2). In view of these results, it is presumed that chimeric RNase enters an intracellular compartment by a receptor-mediated event after which the entire conjugate and/or the RNase alone transits to the cytosol to kill the cell. Cell Biophysics
Volume 21, 1992
128
Rybak et al. Table 2 Specific Reversal of Chimeric RNase Activity Surviving clonogenic units % of Control Tfn-RNase, mM 0 1.4
-Tfn
+Tfn
100 0.002
62 31
Protein synthesis Tfn-RNase
-PRI
+PRI
20
36
100
200
3
100
aAssay protocols described in and data adapted from ref. 13. PRI and Tfn stand for placental RNase inhibitor and transferrin, respectively.
The kinetics of cellular toxicity of plant or bacterial toxins exhibit an initial lag period in which no protein synthesis inhibition is observed (reviewed in ref. 43). Both the length of the lag and the rate of decrease in protein synthesis are concentration-dependent. Similarly, conjugates made with RNase A also exhibit dose-dependent lag times followed by log linear decreases in protein synthesis (13). The rate-limiting step in cytotoxicity of classical immunotoxins is considered to be the translocation of the toxin to the cytosol. Since RNase A directly microinjected into oocytes is roughly 25,000x more potent in abolishing oocyte protein synthesis than RNase A conjugates constructed with chemical linkers, one model would predict that efficient entry of RNase A into the appropriate cellular compartment is hampered. Although bovine pancreatic RNase A is not cytotoxic, a homolog isolated from frogs is a natural toxin (17). Onconase, discussed in the preceeding section, has cytotoxic (51,52) and antitumor properties (19) that appear to be dependent on its RNase activity. Thus it appears that onconase is capable of entering the cytosol. Interestingly and importantly, the ability to enter cells does not result in a more potent conjugate when onconase Cell Biophysics
Volurne 21, 1992
R ibonuclease Chimeras
129
is chemically linked to transferrin or human antitransferrin receptor antibodies (53). These results suggest that optimizing the coupling reaction or the use of different chemical linkers will be necessary to construct more potent chemically linked RNase chimeras.
Recombinant Human RNase Chimeras The lack of sufficient protein has precluded the chemical conjugation of human RNases. Thus it was important to determine whether RNases could be expressed in a functional form as chimeric fusion proteins and whether they would also acquire directed cytotoxic properties when fused to an antibody. The human RNases depicted in Table 1 are both possible candidates for the construction of human RNase chimeras. Based on our studies with nontoxic bovine RNase A compared to the toxic frog enzyme in the construction of RNase chimeras, one might predict that nontoxic human RNases, such as angiogenin, may make fusion proteins with the best therapeutic index. Thus, initial studies were performed with a recombinant form of angiogenin (42) referred to as ANG. The chimeric mouse-human antibody to the transferrin receptor that had been well characterized for its ability to target tumor necrosis factor (44,45) was used to construct ANG fusion proteins. The gene for ANG was modified and cassetted into an expression vector containing the chimeric antibody gene at the 5' region of the CH2 domain of the antibody, thus leaving the hinge region unaffected and dimerization of the heavy chain possible (15). The junctions between the antibody heavy chain and the ANG gene are depicted at the top of Fig. 2. An antitransferrin receptor chimeric lightchain-producing cell line E12B5 (46) was transfected with the vector containing the CH2ANG gene. After selection for the presence of the gpt gene, culture supernatants of clones testing positive for human IgG were followed for reproducible human IgG activity. Of these, clone CH2ANG was selected for further characterization. The amount of secreted CH2ANG ranged from I to 5 ng/mL as determined in a human IgG-detecting ELISA and from I to 2 ng/mL when ANG was detected by ELBA. Unfortunately, the level of secretion of this chimeric antibody linked to ANG or TNF was far lower than that described for other chimeric enzymes (47,48). It is probable that the antibody secretory pathway used by these constructs did not protect the cell from cytotoxicity to the same extent as would the normal bioCell Biophysics
Volume 21, 1992
130
Rybak et al.
Chimeric CH2Ang CK
VL
_l
I VH
CHI
I
H CH2
ANG
I I illlllllllllllllllllllllllllllllllllllllll
I
f
--a-~ntibody ..... CH2- junction ............ ANG........ A
P
GCA CCT
E GAA
F TTC
Q D N S CAG GAT AAC TCC
--EcoRl--
Fig. 2. Sequences around the junction between antibody heavy chain in the CH2 domain and the 5' end of the sequence coding for mature ANG in the CH2ANG construct. E and F are amino acids introduced by PCR modification of the ANG gene prior to cloning. Single-letter amino acid code is used. VH, mouse variable regions of E6 antibody gene; CH1 and CH2, constant regions I and 2, respectively, of the heavy-chain gene; H, hinge region of heavy-chain gene; ANG, DNA coding for human angiogenin is shown in a shaded box. synthetic pathway for these proteins and high-level expression possibly caused the self-selection of low producers. The extremely low yield of protein precluded the purification of the ANG chimera to homogeneity. However, partially purified material was found to express characteristic in vitro activities of the native RNase (Fig. 3A). CH2ANG inhibited protein synthesis in the rabbit reticulocyte lysate, and this was reversed by including an inhibitor of ANG. Importantly, medium containing the fusion protein killed human leukemia cells that express the transferrin receptor (Fig. 3B), while sparing cells that did not Cell Biophysics
Volume 21, 1992
Ribonuclease Chimeras A
131
120
o
100
~
6O
0
r-
~ if)
411
"3
20
o.
0
Control
B
CH2ANG 40 ng
CH2ANG 80 ng
CH2ANG 80 ng+PRI
10001 100
0
5
10
1"5
DAYS IN C U L T U R E
Fig. 3. A: Effect of CH2ANG on protein synthesis in the rabbit reticulocyte lysate in the absence or presence of ANG inhibitor (PRI). B: Growth of K562 cells in transfectoma culture supematant. K562 cells were grown in I mL of growth medium (--B--) or I mL of growth medium that contained 1-2 ng/mL of CH2ANG (me--). The data are adapted from ref. 15. Cell Biophysics
Volume 21, 1992
132
Rybak et al.
recognize the human antibody (15). In addition, cytotoxicity was completely blocked by the daily addition of excess antibody (15). Thus, human RNase can be expressed as a recombinant fusion protein that not only retains some of its individual biological activities, but also acquires cytotoxicity similar to the chemically linked bovine RNase A. Additionally, the construction of a unique linkage between RNase and a cell-binding ligand eliminates the heterogeneity of chemically linked antibody-RNase conjugates. This may have contributed to the increased potency of the recombinant antibody fusion protein. Based on immunoreactivity, the recombinant RNase chimera was about 1000 times more potent than chemically linked RNase hybrids. The established host defense activities of EDN and ECP make them intriguing candidates for immunotherap~ since the concept of targeting natural biological response modifiers is appealing. In this regard, an antibody-tumor necrosis factor hybrid molecule has been expressed and secreted by myeloma cells (44,45), and the TNF fusion protein was found to be more toxic to specific target cells than the native cytokine. For these reasons, studies are ongoing to express RNase as single-chain antibody fusion proteins in E. coli. It is hoped that this methodology will allow the production of sufficient chimeric protein to establish clinical relevance.
In Vivo Antitumor Effects of RNase Chimeras Antitransferrin receptor RNase A conjugates were evaluated for antitumor effects in vivo (14). U251 human glioma tumors were grown in the flanks of nude mice until a tumor size of 0.5-1.0 cm diameter was reached. The tumors were injected intratumorally with four doses of saline, RNase A conjugate, or a nonconjugated mixture of RNase A and the antibody (Fig. 4). Only the RNase A chimera inhibited tumor growth, and the inhibitor effect was maintained up to 30 d when the experiment was terminated. The in vivo effect of the conjugate mimicks in vitro activity in that the unconjugated mixture had no effect on tumor growth. These in vivo data confirm that RNase must be linked to the antibody to acquire cytotoxicity. Although chemically linked RNase chimeras are considerably less active to inhibit in vitro cellular protein synthesis than hybrid proteins constructed with antibodies and classical toxins (13,14), they are sur-
Cell Biophysics
Volume 21, 1992
Ribonuclease Chimeras
133
A
I,LI
1.50
+1 CO
I5 o
1.00 I= "6
E =
l--
0.50
C G)
0.00
0
10
20
30
Days
Fig. 4. Tumoricidal activity of an anitibody-RNase chimera directed against the human transferrin receptor. Solid U251 human glioblastoma tumors were established in the flanks of nude mice. Animals were treated with direct intratumoral injections every other day for a total of four doses. Saline (4, N = 10); RNase mixed with 5E9 antibody (44 and 210 pmol, respectively; m, N = 10), and 5E9 RNase chimera (27-29 pmol; [3, N = 10). Reproduced with permission from (14). prisingly potent in vivo. The antitumor effect of the antitransferrin receptor antibody-RNase A conjugate was equal to that of an antitransferrin receptor ricin A-chain immunotoxin (454A12-rRA) that has been used in a human clinical trial (Laske et al., manuscript in preparation). The disparity in the in vitro and in vivo potency of chemically linked RNase A chimeras hints at factors that may amplify the effect of RNase containing drugs in vivo. If so, RNase chimeras may help to compensate for such problems as antigen heterogeneity, which limits access of immunotherapeutics directed to internalizing antigens.
Cell Biophysics
Volume21, 1992
134
Rybak et al.
SUMMARY Targeting drugs for cancer therapy with antibodies directed to tumor cells offers the hope of selective toxic effects. Targeted therapy with drugs constructed with toxic proteins conjugated to antibodies might offer the best possibility for success since (1) they act catalytically rather than stoichiometrically, and (2) the requirement for cellular internalization of the toxin to kill the target cell helps to assure that nearby normal cells are not killed as they would be with antibodies conjugated to radioisotopes or surface-acting cytolytic agents. In vivo therapy with immunotoxins has been disappointing in the initial trials, but these trials were designed to test treatment-related toxicities and not efficacy. In this regard, many facets of immunotoxin therapy have been discovered that were not predicted from animal models, and experimental protocols to address these problems are being designed. In addition, future advances in immunology coupled to advances in constructing recombinant immunotoxins will contribute to the increasing success of this therapy as a standard treatment of cancer alone or combined with other therapeutic modalities. To date, only nonmammalian toxins have been used to construct immunotoxins, and human therapy with these reagents has been limited by problems of immunogenicity and clinical trial suspending toxicities (6,49,50). Adopting a rational approach toward immunotherapy by using human RNase could minimize two of the major problems identified in clinical trials, thus contributing to the successful application of this form of directed drug therapy.
REFERENCES 1. Gilliland, D. G., Steplewski, Z., Collier, R. J., Mitchell, K, Chang, T., and Koprowski, H. (1980) Antibody directed cytotoxic agents: use of monoclonal antibody to direct the action of toxin A chains to colorectal carcinoma cells. Proc. Natl. Acad. Sci. USA 77, 4539-4543. 2 Krolick, K. A., ViUemez, C., Isakson, P., Urh, J., and Vitetta, E. (1980) Selective killing of neoplastic B cells by antibodies coupled to the A chain of ricin. Proc. Natl. Acad. Sci. USA 77, 5419-5423. 3. Youle, R. J. and Neville, D. M. (1980) Anti-Thy 1.2 monoclonal antibody linked to ricin is a potent cell-type-specific toxin. Proc. Natl. Acad. Sci. USA 77, 5483-5486. Cell Biophysics
Volume 21, 1992
Ribonuclease Chimeras
135
4. Sawler, D. L., Bartholomew, R. M., Smith, L. M., and Dillman, R. (1985) Human immune response to multiple injections of murine monoclonal IgG. J. Immunol. 135,1530-1535. 5. Schroff, R. W., Foon, K. A., Beatty, S. M., Oldham, R., and Morgan, A. (1985) Human anfi-murine immunoglobulin response in patients receiving monoclonal antibody therapy. Cancer Res. 45, 879-885. 6. Rybak, S. M. and Youle, R. J. (1991) Clinical use of immunotoxins: monoclonal antibodies conjugated to protein toxins. Immunol. Allergy Clin. North Am. 11:2, 359-380. 7. Harkonen, S., Stoudemire, J., Mischak, R., Spetler, L., Lopez, H., and Scannon, P. (1987) Toxicity and immunogenicity of monoclonal antimelanoma antibody-ricin A chain immunotoxins in rats. Cancer Res. 47, 1377-1385. 8. Hertler, A. (1988) Human immune response to immunotoxins, in Immunotoxins (Frankel, A. E., ed.), Kluwer Academic, Boston/Dordrecht/ Lancaster, pp. 475-480. 9. Boulianne, G. L., Houzumi, N., and Schulman, M. J. (1985) Production of functional chimeric mouse/human antibodies. Nature (Lond.) 643-646. 10. Morrison, S. L., Johnson, M. J., Herzenberg, L. A., and Oi, V. T. (1984) Chimeric human antibody molecules: mouse antigen-binding domains with human constant regions. Proc. Natl. Acad. Sci. USA 81, 6851-6855. 11. Jones, P. T., Dear, P. H., Foote, J., Neuberger, M. S., and Winter, G. (1986) Replacing the complementary-determining regions in a human antibody with those from a mouse. Nature (London) 321, 522-525. 12. Riechmann, L., Clark, M., Waldmann, H., and Winter, G. (1988) Reshaping human antibodies for therapy. Nature (Lond.) 332, 323-327. 13. Rybak, S. M., Saxena, S. K., Ackerman, E. J., and Youle, R. J. (1991) Cytotoxic potential of RNase and RNase hybrid proteins. J. Biol. Chem. 266, 21,202-21,207. 14. Newton, D. L., Ilercil, O., Laske, D. W., Oldfield, E., Ryback, S. M., and Youle, R. J. (1992) Cytotoxic ribonuclease chimeras: targeted tumoricidal activity in vitro and in vivo. J. Biol. Chem., submitted. 15. Rybak, S. M., Hoogenboom, H. R., Meade, H., Ravs, J. C., Schwartz, D., and Youle, R. J. (1992) Humanization of immunotoxins. Proc. Natl. Acad. Sci. USA 89, 3165-3169. 16. McClure, B. A., Haring, V., Ebert, P. R., Anderson, M. A., Simpson, R. J., Sakiyama, F., and Clarke, A. E. (1989) Style self-incompatibility gene products of Nicotiana alata are ribonucleases. Nature (Lond.) 342, 955-957. Cell Biophysics
Volume 21, 1992
136
R y b a k et al.
17. Darzynkiewicz, Z., Carter, S. P., Mikulski, S. M., Ardelt, W., and Shogen, K. (1988) Cytostatic and cytotoxic effects of Pannon (P-30 protein) a novel anti-cancer agent. Cell Tissue Kinetics 21, 169-182. 18. Mikulski, S. M., Viera, A., Ardelt, W., Menduke, H., and Shogen, K. (1990) Tamoxifen and trifluoroperazine (stelazine) potentiate cytostatic/cytotoxic effects of P-30 protein, a novel protein possessing anti-tumor activity. Cell Tissue Kinetics 23, 237-246. 19. Ardelt, A., Mikulski, S. M., and Shogen, K. (1991) Amino acid sequence of an anti-tumor protein from Rana pipiens oocytes and early embryos. J. Biol. Chem. 256, 245-251. 20. Mikulski, S. M., Ardelt, W., Shogen, K., Vernstein, E. H., and Menduke, H. (1990) Striking increase of survival of mice bearing M109 Madison Carcinoma treated with a novel protein from amphibian embryos. J. Natl. Cancer Inst. 82, 151-153. 21. Beintema, J., Schuller, C., Irie, M., and Carsana, A. (1988) Molecular evolution of the ribonuclease superfamily. Prog. Biophys. Mol. Biol. 51, 165-192. 22. Roth, J. S. (1963) Ribonuclease activity and cancer: a review. Cancer Res. 23, 657-666. 23. Matousek, J. (1973) The effect of bovine seminal ribonuclease (AS RNase) on cells of Crocker tumour in mice. Experientia 29, 858,859. 24. Vescia, S., Tramontano, D., Augusti-Tocco, G., and D'Allessio, G. (1980) In vitro studies on selective inhibition of tumor cell growth by seminal ribonuclease. Cancer Res. 40, 3740-3744. 25. Slifman, N. R., Loegering, D. A., McKean, D. J., and Gleich, G. J. (1986) Ribonuclease activity associated with human eosinophil-derived neurotoxin and eosinophil cationic protein. J. Immunol. 137, 2913-2917. 26. Spry, C. (1988) Eosinophils: A Comprehensive Review and Guide to the Scientific and Medical Literature. Oxford University Press, Oxford, New York. 27. Tepper, R. I., Pattengale, P. K., and Leder, P. (1989) Murine interleukin4 displays potent anti-tumor activity in vivo. Cell 57, 503-512. 28. Ghiabi, M., Gallagher, G. T., and Wong, D. T. (1992) Eosinophils, tissue eosinophilia and eosinophil-derived transforming growth factor A in hamster oral carcinogenesis. Cancer Res. 52, 389-893. 29. Young, J., Peterson, C., Venge, P., and Cohn, Z. A. (1986) Mechanism of membrane damage mediated by human eosinophil cationic protein. Nature (Lond.) 321, 613-616. 30. Rosenberg, H. F., Tenen, D. G., and Ackerman, S. J. (1989) Molecular cloning of the human eosinophil-derived neurotoxin: a member of the ribonuclease gene family. Proc. Natl. Acad. Sci. USA 86, 4460--4464. Cell Biophysics
Volume 21, 1992
Ribonuclease Chimeras
137
31. Reddi, E. K. (1975) Nature and possible significance of human serum ribonuclease. Biochem. Biophys. Res. Commun. 67, 110-118. 32. Blank, A., Dekker, C., Schieven, G., Sugiyama, R., and Thelen, M. (1981) Human Body Fluid Ribonucleases: Detection, Interrelationships and Significance. IRL, London, pp. 203-209. 33. Fett, J. W., Strydom, D. J., Lobb, R. R., Alderman, E. M., Bethune, J. L Riordan, J. F., and Vallee, B. L. (1985) Isolation and characterization of angiogenin an angiogenic protein from human carcinoma cells. Biochemistry 24, 5480-5485. 34. Strydom, D. J., Fett, J. W., Lobb, R. R., Alderman, E. M., Bethune, J. L Riordan, J. F., and Vallee, B. L. (1985) Amino acid sequence of human tumor derived angiogenin. Biochemistry 24, 5486-5494. 35. Kurachi, K., Davie, E. W., Strydom, D. J., Riordan, J. F., and Vallee, B. L. (1985) Sequence of the cDNA and gene for angiogenin, a human angiogenesis factor. Biochemistry 24, 5494-5499. 36. Shapiro, R., Riordan, J. F., and Vallee, B. L. (1986) Characteristic ribonucleolytic activity of human angiogenin. Biochemistry 25, 3527-3531. 37. St. Clair, D. K., Rybak, S. M., Riordan, J. F., and Vallee, B. L.(1987) Angiogenin abolishes cell-free protein synthesis by specific ribonucleolytic inactivation of ribosomes. Proc. Natl. Acad. Sci. USA 84, 8330-8334. 38. O'Keefe, D. O. and Draper, R. K. (1986) Characterization of a transfertin-diphtheria toxin conjugate. J. Biol. Chem. 260, 932-937. 39 Raso, V. and Basala, M. (1984) A highly cytotoxic human transferrinricin A chain conjugate used to select receptor-modified cells. J. Biol. Chem. 259, 1143-1149. 40. Wagner, E., Zenke, M., Cotten, M., et al. (1990) Transferrin-polycation conjugates as carriers for DNA uptake into cells. Proc. Natl. Acad. Sci. USA 87, 3410-3414. 41. Prior, T. I., Fitzgerald, D. J., and Pastan, I. (1991) Bamase Toxin: a new chimeric toxin composed of pseudomonas exotoxin A and Barnase. Cell 64, 1017-1023. 42. Saxena, S. K., Rybak, S. M., Winkler, G., Meade, H. M., Mc Gray, P., Youle, R. J., and Ackerman, E. J. (1991) Comparison of RNases and toxins on injection into Xenopus oocytes. J. Biol. Chem. 266, 21,20821,214. 43. Hudson, T. H. and Neville, D. M. (1988) Enhancement of immunotoxin action: Manipulation of the cellular routing of proteins, in Immunotoxins (Frankel, A. E., ed.), Kluwar Academic, Boston/Dordrecht/Lancaster, pp. 371-389. Celt Biophysics
Voturne 21, 1992
138
R y b a k et al.
44. Hoogenboom, H. R., Raus, J. M., and Volckaert, G. (1991) Targeting of tumor necrosis factor to tumor cells: secretion by myeloma cells of a genetically engineered antibody-tumor necrosis factor hybrid molecule. Biochem. Biophys. Acta 1096, 345-354. 45. Hoogenboom, H. R., Volckaert, G., and Raus, J. M. (1991) Construction and expression of antibody-tumor necrosis factor fusion proteins. Mol. Immunol. 28, 1027-1037. 46. Hoogenboorn, H. R., Raus, J. M., and Volckaert, G. (1990) Cloning and expression of a chimeric antibody directed against the human transferrin receptor. J. Immunol. 144, 3211-3217. 47. Casadei, J., Powell, M. J., and Kenten, J. H. (1990) Expression and secretion of aequorin as a chimeric antibody by means of a mammalian expression vector. Proc. Natl. Acad. Sci. USA 87, 2047-2051. 48. Williams, G. T. and Neuberger, M. S. (1986) Production of antibody tagged enzymes by myeloma cells: application to DNA polymerase I Klenow fragment. Gene 43, 319-324. 49. Pastan, I., Willingham, M., and FitzGerald, D. (1986) Immunotoxins. Cell 47, 641-648. 50. Vitetta, E. S., Fulton, R. J., and May, R. D. (1987) Redesigning nature's poisons to create anti-tumor reagents. Science 238, 1098-1104. 51. Newton, D. L., Walbridge, S., Mikulski, S. M., Ardelt, K. S., Ackerman, S. J., Rybak, S. M., and Youle, R. J. (1983) Toxicity of an antitumor ribonuclease to Purkinje neurons. J. Neuroscience, in press. 52. Wu, Y. N., Mikulski, S. M., Ardelt, W., Rybak, S. R., and Youle, R. J. (1993) A cytotoxic ribonuclease. J. Biol. Chem. 268, 10,686-10,693. 53. Rybak, S. M., Newton, D. L., Mikulski, S. M., Viera, A., and Youle, R. J. (1993) Cytotoxic onconase and ribonuclease A chimeras. Drug Targeting and Delivery, in press.
Cell Biophysics
Volume 21, 1992
