Use of Immunotoxin to Inhibit Proliferating Human Corneal Endothelium

Samuel Fulcher,* Geming Lui,f L. L. Houston,£ 5. Romokrishnon,^: Terry Burris,§ Jon Polansky,t ond Jorge Alvaradof

Transferrin plays a central role in cellular proliferation and proliferating ceils have been shown to express transferrin receptors with increased density. We examined the effect of an immunotoxin consisting of anti-transferrin receptor monoclonal antibody (454A12) conjugated to recombinant ricin A chain (rRTA) on the proliferation of human corneal endothelium (HCE) in vitro. In proliferating cultures an immunotoxin (454A12-rRTA) concentration of 50 ng/mL reduced cell counts at day 7 by at least 89%, with no effect observed at 0.01 ng/ml. In contrast, cell counts were only minimally reduced in confluent cultures, even after 7 days' exposure to high concentrations of immunotoxin. These data suggest that 454A12-rRTA may be used to prevent growth of human corneal endothelium in patholog- ical conditions such as the iridocorneal endothelial (ICE) syndrome. Invest Ophthalmol Vis Sci 29:755-759,1988

The hyperproliferation of nonmalignant tissue munotoxins is the development of monoclonal anti- causes a wide spectrum of clinically significant ocular bodies that specifically react with antigens on the sur- disease such as proliferative vitreoretinopathy, the ir- faces of cells and that can be used to carry potent idocorneal endothelial syndromes, fibrous or epithe- cellular toxins to target cells. Several toxins or frag- lial downgrowth and posterior capsular fibrosis. ments of toxins such as diphtheria toxin, ricin and Methods currently used to treat these conditions in- ricin A chain have been coupled chemically to anti- clude surgery, laser surgery, cryotherapy and chemo- bodies to form immunotoxins.12 Ricin A chain is an therapy using 5-fluorouracil. These methods often enzyme that inhibits protein synthesis by chemically fail completely or are unsatisfactory because they do modifying 28 S rRNA on the 60 S ribosomal subunit not successfully differentiate between unwanted, pro- to prevent protein synthesis.3 Because of the presence liferating cells from normal tissue. of the antibody in the conjugate, immunotoxins rec- The use of immunotoxins offers a way to selec- ognize the surface of specific cells. If internalization tively deliver drugs to proliferating cells for treatment occurs, the target cell may be killed selectively. We of some ocular diseases. Central to the use of im- tested the ability of such an immunotoxin directed against transferrin receptor to selectively inhibit the growth of proliferating human corneal endothelial From the *Scott and White Clinic, Scott and White Memorial Hospital, Scott, Sherwood and Brindley Foundation, Texas A&M cells in vitro. University College of Medicine, Temple, Texas, the fUniversity of Transferrin is a growth factor that is required for California Medical Center, San Francisco, California, the |CETUS the growth of all cells, and through endocytosis using Corporation, Emeryville, California, and the §Departments of the transferrin receptor contributes to the maximal Clinical Investigation and Ophthalmology, U.S. Naval Hospital, 45 Oakland, California. proliferation of cells in vitro. Other growth factors Work performed at the U.S. Naval Hospital, Oakland, was such as erythropoietin and epidermal growth factor sponsored under the United States Navy Clinical Investigation Pro- influence cell proliferation and the expression of gram, Study Number 85-48-2098. Work sponsored at the Univer- transferrin receptors.6'7 Because transferrin receptor sity of California, San Francisco, was supported by grants is located at the cell surface, is present in increased EY-02068, EY-02477, and EY-02162. amounts on the surface of proliferating cells,8"10 and The opinions or assertions expressed herein are those of the 810 authors and are not to be construed as official or as necessarily is rapidly internalized after it binds to transferrin, reflecting the views of the U.S. Department of the Navy or the it has been an effective target for the inhibition of Naval Service at large. malignant cells by immunotoxins.1'211"13 Because ac- Submitted for publication: March 31, 1987; accepted December tively proliferating cells express a higher number of 4, 1987. transferrin receptors, we think that proliferating cells Reprint requests: Samuel Fulcher, MD, Department of Ophthal- mology, Scott and White Clinic, Scott and White Hospital, 2401 may be killed preferentially over resting cells by an South 31st Street, Temple, TX 76508. immunotoxin directed toward transferrin receptors.

755

Downloaded from iovs.arvojournals.org on 09/26/2021 756 INVESTIGATIVE OPHTHALMOLOGY 6 VISUAL SCIENCE / May 1988 Vol. 29

In this study, we show that an immunotoxin that Assay of Immunotoxin Activity contains recombinant ricin A chain and an antibody The effect of 454A12-rRTA on cellular prolifera- directed against transferrin receptor selectively in- tion was assayed using two sets of duplicate plates hibits proliferation of dividing human corneal endo- containing serial dilutions of immunotoxin. Four sets thelial cells compared to confluent cell cultures. of controls were used, including plates with unconjugated rRTA, with purified unconjugated 454A12 an- Materials and Methods tibody, with rRTA conjugated to MOPC21 (an irrelevant immunotoxin) or without any additives. Prolif- Production and Characterization erating cells were exposed continuously to a single of Monoclonal Antibodies dose of 454A12-rRTA or control protein for all 7 Murine monoclonal antibody 454A12 (IgG2a iso- days of culture, whereas confluent cultures were ex- tope) is specific for human transferrin receptor and posed for 7 days after achieving confluence. Dead was produced by standard fusion techniques de- cells were observed to detach from the dishes and scribed previously.2 This antibody binds to cell sur- were aspirated prior to counting; unexposed plates face transferrin receptors, immunoprecipitates the did not contain detached nonviable cells. No de- 95K. MW subunit transferrin receptor, recognizes an tached cells were viable as assessed by trypan blue antigenic site distant to the actual transferrin binding exclusion. The number of live cells was determined site and does not interfere with transferrin binding to with a Coulter counter after 7 days of incubation with its receptor or to the internalization of transferrin. immunotoxin or control protein. Ammonium sulfate precipitation, DEAE-Sepharose The effect of immunotoxin on amino acid incorpo- chromatography, and Sephacryl S-300 gel filtration ration was assayed in duplicate with 3 X 104 cells in were used to purify the antibody contained in ascites 0.5 mL of medium seeded in borosilicate glass vials fluid from Balb/c mice. (ICN Radiochemicals, Irvine, CA) coated with 0.2% gelatin. After 48 hr of proliferation, cells were incu- Immunotoxin Synthesis bated in duplicate for 18 hr in the presence or absence of 454A12-rRTA. The cells were then rinsed three Recombinant ricin A chain was obtained from E. times with phosphate-buffered saline (PBS) and incu- co/i and purified to over 99% homogeneity. Imino- bated with 0.2 ml of leucine-free 1640 RPMI thiolane was used to introduce a thiol group onto the (GIBCO) containing 2.0 ^Ci of [H3H]leucine (Du- surface of 454A12 and a mixed disulfide bond was pont NEN Research Products, Boston, MA). Fetal made using Ellman's reagent [5,5'-dithiobis(2-nitro- calf serum (10%) was added to maintain cellular ad- benzoic acid), DTNB]. Disulfide exchange between herence during the incubation with [H3H]leucine. the mixed disulfide bond and the free thiol group of After 3 hr of incubation, the cells were washed three rRTA formed a disulfide bond between 454A12 and times with PBS and the protein was precipitated with rRTA. The immunotoxin was purified by a combina- 5% trichloroacetic acid (GIBCO). Counts were mea- tion of chromatography and gel filtration; the final sured in a Tricarb® liquid spectrometer scintillation product contained no detectable unconjugated rRTA counter with 5 ml of scintillation fluid (Research or454A12. Products International, Elk Grove, IL).

Cell Cultures Data Analysis Under sterile conditions, human corneal endothe- The effects of 454A12-rRTA on cellular prolifera- lial cells were scraped from donor corneas within 2 hr tion and [H3H]Ieucine incorporation were deter- of harvesting. Stock cultures were seeded on gelati- mined by comparing the growth or incorporation in nized 35 mm dishes (Becton/Dickenson, Oxnard, plates or vials containing immunotoxin with control CA) in medium 199 (GIBCO, Grand Island, NY) plates or vials from the same stock to which no containing 15% fetal calf serum (Sterile Systems, 454A12-rRTA was added. Results from plates or Logan, UT), Earle's balanced salts (GIBCO), and 1% vials without additives were defined as representing of 200 mM glutamine (Sigma, St. Louis, MO). The 100% of control numbers for that experiment and cells were maintained in 5% CO2 at 37°C with fibro- were compared to results from plates or vials with blast growth factor (University of California, San 454A12-rRTA, rRTA or 454A12 alone. Percent cell Francisco) added every other day. Once confluent, survival and [H3H]leucine incorporation were calcu- cells were trypsinized and seeded in 2 ml of medium lated for plates or vials containing additives as com- on gelatinized 35 mm dishes at 2 X 104 cells/ml for pared to negative controls. These data were plotted individual experiments. against the log dose (ng/ml) of 454A12-rRTA.

Downloaded from iovs.arvojournals.org on 09/26/2021 No. 5 IMMUNOTOXIN INHIBITS HUMAN CORNEAL ENDOTHELIUM / Fulcher er ol. 757

6 100

10'' 10° 10' 10' IMMUNOTOXIN CONCENTRATION (ng/ml) IMMUNOTOXIN CONCENTRATION (ng/ml) Fig. 1. Percentage of surviving cells recovered as compared to Fig. 2. Percentage of surviving cells in plates with 454A12-rRTA controls plotted against immunotoxin level in confluent and prolif- (solid squares) is lower than in plates with 454A 12-rRTA and 2500 erating HCE cells. Proliferating HCE cells (solid squares) are signif- ng/mL of added 454A12 (open squares). icantly inhibited by 454A12-rRTA while confluent HCE cells are relatively unaffected (open squares). Unconjugated antibody 454A12 (solid circles) and unconjugated ricin A chain (solid trian- million cells) had no effect on confluent cells. HCE gles) do not inhibit proliferating HCE cells. cells in confluent cultures did not become dysmorphic after exposure to 454A12-rRTA; however, proliferating HCE cells quickly became dysmorphic. Results The stability of the immunotoxin during the course Inhibition of Proliferation With Immunotoxin of the study was demonstrated by removing samples of the culture medium and assaying for the presence Proliferating HCE cells were significantly inhibited of unbound immunotoxin on a sensitive cell line by 454A12-rRTA in a dose-dependent fashion (Fig. (MCF-7). There was no change in specific activity of 1). Addition of 454A12-rRTA at 50 ng/ml resulted in unbound immunotoxin over the course of the exper- at least an 89% decrease in the number of viable cells iment. compared with controls without 454A12-rRTA, and the few remaining live cells were dysmorphic on light Inhibition of Protein Synthesis microscopy. Little or no effect was observed at 0.01 Protein synthesis, as measured by the incorpora- ng/ml. Cell survival at an immunotoxin concentra- 3 tion of 50 ng/ml ranged from 50% to 75% of the tion of [H H]leucine into protein, was significantly original number plated. inhibited in a dose-dependent fashion by 454A 12- Unconjugated rRTA or 454A12 alone failed to di- rRTA in both proliferating and confluent HCE cells minish cell numbers (Fig. 1). The irrelevant immu- compared to unexposed controls (Fig. 3). The average notoxin, MOPC21-rRTA, diminished cell counts cell density for confluent cells was four times that of compared with controls by 25% at a concentration of 250 ng/ml, presumably through nonspecific binding and internalization (data not shown). In a separate experiment, the effect of the immunotoxin was shown to be specific since its action could be readily blocked by the presence of unconjugated 454A12 at a concentration of 2500 ng/ml (Fig. 2). In contrast to the relative sensitivity of proliferating HCE cells, confluent cultures were much less affected by 454A12-rRTA, as shown by measuring the number of live cells after treatment with various concentrations of immunotoxin (Fig. 1). There was no significant drug effect up to a concentration of 2500

ng/ml. At 50 ng/ml, a concentration that completely IMMUNOTOXIN CONCENTRATION (ng/ml) inhibited the growth of proliferating cells, the ratio of immunotoxin to cell number was 6.95 pmol per one Fig. 3. Percentage protein synthesis for confluent and proliferating HCE as compared to unexposed controls measured by [H3H]- million cells. A concentration of immunotoxin 50 leucine incorporation. The dose dependency for proliferating (solid times higher (2.5 Mg/ml, equivalent to 28.7 pmol per markers) and confluent HCE cells (open markers) is apparent.

Downloaded from iovs.arvojournals.org on 09/26/2021 758 INVESTIGATIVE OPHTHALMOLOGY b VISUAL SCIENCE / Moy 1988 Vol. 29

proliferating cells during [H3H]leucine incorporation ceptors include basal epidermis, hepatocytes, Kupffer studies. Although cells grown at either density were cells, pituitary, testis and endocrine pancreas.17 Tis- inhibited, there was a differential sensitivity to the sues that would also be theoretically at risk to immunotoxin. After correcting for the number of 454A12-rRTA include those with rapid rates of turn- cells by using the ratio of immunotoxin to one mil- over, such as hematopoietic cells or cells lining the lion cells, proliferating cells (2.32 pmol per one mil- gastrointestinal tract. However, for many regional lion cells) were estimated to be at least 2.5-fold more applications such as in the eye, the amount of im- sensitive than confluent cells (5.8 pmol per one mil- munotoxin administered would be small and should lion cells). This ratio was calculated from the concen- not reach significant levels systemically. The ocular trations of immunotoxin that inhibited 50% of the condition that could benefit from the immunotoxin protein synthesis, 100 ng/ml and 1000 ng/ml for pro- must be defined precisely, with consideration given to liferating and confluent cells, respectively. the dose, method of administration and results of toxicity studies. In vivo studies are needed to better Discussion address these issues prior to the clinical use of immunotoxins. An increase in the expression of transferrin recep- The best measure of cell death in vitro is obtained tors occurs with cell division in both hemoglobin- from counting the number of living cells, and this and non-hemoglobin-producing cells, as described by 57 91415 assay clearly shows differential sensitivity between many investigators. - Immunotoxins directed 1 confluent and proliferating cells. These data derived against the transferrin receptor " or transferrin con- 12 from cell counts are encouraging, but it appears from jugated to rRTA are efficient inhibitors of cellular the [H3H]leucine data that functional transferrin re- proliferation. Most attention has been directed to the ceptor activity in confluent HCE cells is not com- use of such conjugates to inhibit malignant cells, and pletely eliminated. In this case the inhibition of pro- little work has been done to investigate the possibility tein synthesis observed in the confluent HCE cells of using these conjugates to inhibit the hyperprolifer- may represent a nonlethal insult to the resting cell. ation of normal tissue. Although one can postulate that a small proliferating Hyperproliferation of ocular cell types occurs in a subpopulation of HCE cells in the confluent cultures wide spectrum of ocular diseases, including prolifera- accounts for the decrement in [H3H]leucine incorpo- tive vitreoretinopathy, epithelial downgrowth, the ration, this appears unlikely since baseline protein ICE syndromes, recurrent pterygium, filtration sur- synthesis per cell was not diminished in confluent gery failure and posterior capsular fibrosis after ex- HCE cell controls compared with proliferating HCE tracapsular cataract extraction. Methods currently controls. used to treat these conditions include surgery, 5-fluo- Although these data suggest that selectivity for pro- rouracil, cryotherapy and laser applications. By the liferating HCE cells can be achieved in vitro, it is not selective inhibition of proliferating cells with specific clear that similar results will be obtained in vivo. It is immunotoxins, treatment of some of these diseases clear that immunotoxins directed against transferrin may be possible. receptors are capable of preventing tumor cell growth Alvarado et al16 have recently discussed the estab- in vivo, and that recombinant ricin A chain can be lishment of long-term culture techniques for HCE 13 substituted for ricin A chain obtained from plants. cells. We chose HCE cells as the target tissue because It would be helpful to study the effects of 454A12- of the potential application of immunotoxin in the rRTA or similar immunotoxins in suitable in vivo ICE syndrome and because of our experience in models of ocular tissue hyperproliferation, or in in working with these cells. Our data clearly demon- vitro models of wound healing to better ascertain the strate that proliferating HCE cells may be selectively potential application of such agents in the treatment killed by 454A12-rRTA as compared to confluent of clinical ocular disease. Further investigation into HCE cells. This suggests that it may be possible to the role of transferrin and transferrin receptors in oc- treat hyperproliferative eye disorders in vivo without ular cell division may provide new avenues to pro- significantly damaging surrounding eye tissue. In this mote or reduce cell proliferation in the eye. Our find- regard it is important to show that whether sur- ings with proliferating and confluent HCE cells sug- rounding eye tissues do not express active transferrin gest the possible eventual therapeutic use of receptors in significant amounts or are otherwise re- immunotoxins in hyperproliferative conditions of sistant to the detrimental effects of the immunotoxin. the eye. Normal human tissues express transferrin receptors in varying density at rest or during cell division.17 Key words: immunotoxin, cell proliferation, corneal endo- Tissues that express higher densities of transferrin re- thelium, transferrin, transferrin receptor

Downloaded from iovs.arvojournals.org on 09/26/2021 No. 5 IMMUNOTOXIN INHIBITS HUMAN CORNEAL ENDOTHELIUM / Fulcher er ol. 759

Acknowledgments mediated endocytosis of human transferrin receptor and its cell surface receptor. J Cell Physiol 124:283, 1985. The authors wish to thank Mary Gavin, PhD, and Deb- 9. Frazier JL, Caskey JH, Yoffe M, and Seligman PA: Studies of bie Anderson for their invaluable assistance in the statistical the transferrin receptor on both human reticulocytes and nu- analysis of this study. Additionally, we express our appreci- cleated human cells in culture. J Clin Invest 69:853, 1982. ation to Scott Hambly, Dana Hockett Evans and Joan 10. Larrick JW and Cresswell P: Modulation of cell surface iron Carter for their editorial assistance in preparation of this transferrin receptors by cellular density and state of activation. manuscript. J Supramol Struc 11:579, 1979. 11. Ramakrishnan S and Houston LL: Inhibition of human acute References lymphoblastic leukemia cells by immunotoxin: Potentiation by chloroquine. Science 223:58, 1984. 1. Trowbridge IS and Domingo DL: Anti-transferrin receptor 12. Raso V and Basala M: A highly cytotoxic human transferrin- monoclonal antibody and toxin antibody conjugates affect ricin A chain conjugate used to select receptor modified cells. J growth of human tumour cells. Nature 294:171, 1981. Biol Chem 259:1143, 1984. 2. Bjom MJ, Ring D, and Frankel A: Evaluation of monoclonal 13. Fitzgerald DJ, Bjorn MJ, Ferris RJ, Winkelhabe JL, Frankel antibodies for the development of breast cancer immuno- AE, Hamilton TC, Ozols RF, Willingham MC, and Paston J: toxins. Cancer Res 45:1214, 1985. Antitumor activity of two immunotoxins directed against 3. Endo Y, Mitsui K, Motiyuki M, and Tsurugi K: The mecha- human ovarian cancer cells in nude mice. Cancer Res 47:1407, nism of action of ricin and related toxic lectins on eukaryotic 1987. ribosomes. J Biol Chem 262:5908, 1987. 14. Trowbridge IS and Omary BM: Human cell surface glycopro- 4. Barnes DJ and Sato G: Serum-free cell culture: A unifying tein related to cell proliferation in the receptor for transferrin. approach. Cell 22:649, 1980. Proc Natl Acad Sci USA 79:3039, 1981. 5. Seligman PA: Structure and function of the transferrin recep- 15. Chitambar CR, Maskey EJ, and Seligman PA: Regulation of tor. ProgHematol 13:131, 1983. transferrin receptor expression on human leukemic cells dur- 6. Shannon KM, Larrick JW, Fulcher SA, Burck KB, Weintraub ing proliferation and induction of differentiation. J Clin Invest H, Davis JC, and Ring D: Selective inhibition of human ery- 22:1314, 1983. throid burst forming units by monoclonal antibodies against 16. Alvarado J, Lui GM, and Wood I: Establishment of human transferrin or the transferrin receptor. Blood 67:1631, 1986. corneal endothelial cells in tissue culture. ARVO Abstracts. 7. Wiley SH and Kaplan J: Epidermal growth factor rapidly in- Invest Ophthalmol Vis Sci 27(Suppl):3, 1986. duces a redistribution of transferrin receptor pools in human 17. Gatter KC, Brown G, Trowbridge IS, Woolston R, and Mason fibroblasts. Proc Natl Acad Sci USA 81:7456, 1984. DY: Transferrin receptors in human tissues: Their distribution 8. Larrick J, Enns C, Raubitshek A, and Weintraub H: Receptor and possible clinical relevance. J Clin Pathol 36:539, 1983.

Downloaded from iovs.arvojournals.org on 09/26/2021