INFECTION AND IMMUNITY, Apr. 1975, p. 675-684 Copyright 0 1975 American Society for Microbiology

Vol. 11, No. 4 Printed in U.S.A.

Interaction of Cultured Mammalian Cells with [125I]Diphtheria Toxin PETER F. BONVENTRE,* CATHARINE B. SAELINGER, BRUCE IVINS, CASIMIR WOSCINSKI, AND MICHAEL AMORINI Department of Microbiology, University of Cincinnati, College of Medicine, Cincinnati, Ohio 45267 Received for publication 2 December 1974

The characteristics of cell adsorption and pinocytotic uptake of diphtheria toxin by several mammalian cell types were studied. Purified toxin iodinated by a solid-state lactoperoxidase method provided preparations of high specific activity and unaltered biological activity. Diphtheria toxin-sensitive HEp-2 cells and guinea pig macrophage cultures were compared with resistant mouse L-929 cells. At 37 C the resistant cells in monolayer adsorbed and internalized [1251Itoxin to a greater extent than did the HEp-2 cell cultures; no significant differences were observed at 5 C. Ammonium chloride at protective levels did not alter uptake of toxin by either L-929 or HEp-2 cells. Biological activity of the iodinated toxin, however, was negated provided the presence of ammonium chloride was maintained. The ammonium salt appears to maintain toxin in a state amenable to antitoxin neutralization. Guinea pig macrophages internalized iodinated toxin to a level 10 times greater than the established cell lines. In spite of the increased uptake of toxin by the endocytic cells, ammonium chloride prevented expression of toxicity. In an artificial system, toxin adsorbed to polystyrene latex spheres and internalized by guinea pig macrophages during phagocytosis did express biological activity. Ammonium chloride afforded some but not total protection against toxin present in the phagocytic vacuoles. The data suggest that two mechanisms of toxin uptake by susceptible cells may be operative. Toxin taken into the cell by a pinocytotic process probably is not ordinarily of physiological significance since it is usually degraded by lysosomal enzymes before it can reach cytoplasmic constituents on which it acts. When large quantities of toxin are pinocytized, -toxicity may be expressed before enzymatic degradation is complete. A more specific uptake involving direct passage of the toxin through the plasma membrane may be the mechanism leading to cell death in the majority of instances.

The fundamental differences accounting for expression of diphtheria toxin action. Several of susceptibility or resistance to diphtheria toxin these compounds, such as sodium arsenite, are are largely undefined. During the past several metabolic poisons, but another, ammonium years, however, due primarily to the elucidation chloride, is a relatively nontoxic compound used of a molecular basis of toxicity (18), it has been routinely in biochemical reaction mixtures. We possible to initiate an analysis of the problem. have substantiated by indirect means that amSpecies resistance to toxin varies by three to monium chloride does not prevent the initial four orders of magnitude. For example, hu- interaction of toxin with the cell surface, a mans, canines, and guinea pigs are extremely process that may involve either adsorption via sensitive species, whereas rats and mice are electrostatic forces (13) and/or specific receptor relatively resistant (1). Since cell-free protein- sites for toxin on the plasma membrane (21). synthesizing systems from all mammalian spe- Furthermore, we observed that ammonium cies tested are equally sensitive to the direct chloride has no direct toxin-neutralizing effect addition of toxin (7, 23), it is clear that natural per se. Since we also established (22) that the resistance depends upon cellular integrity and it salt does not inhibit toxin-mediated catalytic is likely that the components and mechanisms inactivation of cytoplasmic elongation factor 2 responsible for toxin susceptibility or resistance (11), it was logical to determine whether ammoare located on or in the plasma membrane. In a nium chloride modified the movement of toxin companion paper (22) we documented that across the cell membrane. The classical method several simple inorganic salts inhibit or abolish of uptake of proteins by mammalian cells is by a (675

676

BONVENTRE ET AL.

pinocytic mechanism (33), and since this process may be monitored by quantitating the uptake of iodinated proteins, we conducted experiments that followed the uptake of [125I]_ diphtheria toxin by a diphtheria-sensitive (HEp-2) and a diphtheria-resistant (L-929) cell in the presence and absence of protective levels of ammonium chloride. The data show that both sensitive and resistant mammalian cells adsorb and internalize the protein toxin in significant amounts. They also show that diphtheria toxin within endocytic vesicles may or may not express toxicity, depending upon environmental conditions. Evidence is presented suggesting that toxin taken into susceptible cells by pinocytosis is usually not expressed. Lastly, the experiments show that ammonium chloride has no effect on pinocytic uptake of toxin although its continued presence prevents expression of any toxicity. MATERIALS AND METHODS Toxin. Preparation of toxin and antitoxin were the same as described in the preceding paper (22). The biological activity of toxin preparations, before and after iodination, was tested by guinea pig assay (2). Serial dilutions of toxin (0.5 ml in phosphate-buffered saline, pH 6.9) were injected intracardially and time to death was determined. The minimum lethal dose was calculated from a standard curve. Cells. HEp-2 cells were maintained and experimental cultures established as previously described (22). Guinea pig macrophages were harvested from the stimulated peritoneum and cultured as described in detail elsewhere (22). Strain L, NCTC clone 929, obtained from the American Type Culture Collection, Rockville, Md., is a fibroblast-like cell of mouse connective tissue origin (ATCC CCL 1.2). These cells were maintained as monolayers in NCTC-135 medium (GIBCO, Grand Island, N.Y.) containing 10% horse serum (GIBCO) and the antibiotics penicillin (200 U/ml) and streptomycin (0.2 mg/ml). Before use, L-929 cells were seeded into plastic dishes (15 by 60 mm; Falcon Plastics, Los Angeles, Calif.) at a concentration of 106 cells per dish and incubated for 24 h at 37 C in an environment of 5% CO2. lodination procedures. Two methods of iodination were used. In initial experiments, proteins were iodinated by the chloramine T method of Greenwood et al. (19) as modified by Pappenheimer and Brown (29). The standard reaction mixture contained 5 or 10 mCi of Na'251 (New England Nuclear Corp., Boston, Mass.), 200 Ag of chloramine T, and 1 mg of protein (diphtheria toxin or bovine serum albumin [BSA]) in sodium phosphate buffer (0.5 M, pH 7.6) in a total volume of 0.54 ml. The iodination reaction was allowed to proceed for 6 min at room temperature and was stopped by addition of 0.1 ml of sodium metabisulfite (240 gg). Then 0.4 ml of 1.0% potassium iodide solution in buffer containing 0.05% bovine serum BSA was added, and labeled protein was separated from free iodine and protein aggregates by passage through

INFECT. IMMUN.

a Sephadex G-100 column. The column was equilibrated and eluted with 0.01 M potassium phosphate buffer, pH 6.9, containing 0.05% BSA. The iodinated protein was used immediately after preparation in

uptake experiments. Since the previously described iodination procedure uses a relatively high concentration of the strong oxidizing agent chloramine T, significant loss of biological activity may occur. For later experiments, therefore, diphtheria toxin was iodinated by the solid-state lactoperoxidase procedure of David (12). This enzymatic method of radio-iodinating proteins utilizes bovine lactoperoxidase coupled to cyanogen bromide-activated Sepharose 4B (Pharmacia, Uppsala, Sweden). Activated Sepharose 4B covalently linked with lactoperoxidase was either purchased from Worthington Biochemical Corp., Freehold, N.J.,

or was prepared by the method of David (12). lodinations were carried out at room temperature for 30 min with constant agitation. The reaction mixture contained: 5 mCi of Na'25I (50 Ml), Sepharose 4B activated beads containing approximately 5 U of lactoperoxidase (250 ,l), 1 to 2 mg of diphtheria toxin (250 gl), 10 ul of 30% H202, and sufficient sodium phosphate buffer (pH 6.5, 0.25 M) to bring the final volume to 0.76 ml. The labeled toxin was separated from free Na125I as described for the chloramine T

procedure.

Procedures for measuring uptake of iodinated protein. The following procedure was used in the initial experiments for measuring uptake of iodinated toxin or BSA by HEp-2 and L-929 cells. Appropriate tissue culture medium containing 2 x 106 to 4 x 106 counts/min per ml of freshly labeled [125I]protein (chloramine T method) was incubated with cell monolayers for up to 3 h. At the indicated times, the monolayers were washed four times with warm Hanks balanced salt solution (HBSS). Cold HBSS then was added to the monolayers, and the cells were scraped loose with a rubber policeman, pipetted into centrifuge tubes, pelleted, and resuspended in 5 ml of cold

distilled water. Sufficient trichloroacetic acid was added to each tube to obtain a final concentration of 6% acid. After overnight storage at 4 C, the precipitates were collected on polyvinyl chloride filters (Millipore Corp., New Bedford, Mass.) and washed with cold 6% trichloroacetic acid. The filters were dried, put into counting tubes, and assayed for radioactivity in a well gamma counter (Packard Corp., LaGrange, Ill.). After counting, the filters were incubated with 2 ml of Lowry reagent C (26) for 3 h at 80 C to digest the precipitate. Aliquots of the digests were assayed for protein content by the method of Oyama and Eagle (28). Adsorption and uptake are expressed as counts per min of ['251]toxin per milligram of protein. After the initial experiments, the protocol for measuring uptake of radioactive protein was modified to reduce the level of nonspecific radioactivity associated with the cells. Monolayers were incubated at 4 or 37 C with tissue culture medium containing between 106 and 5 x 106 counts/min of [125Ildiphtheria toxin per ml. After an initial 5-min incubation, a zero-time sample was taken to determine the level of back-

VOL. 11, 1975

677

MAMMALIAN CELLS AND [1251 ]DIPHTHERIA TOXIN

ground adsorption. Incubation was continued and samples were taken at indicated intervals and processed as follows: monolayers were drained, washed three times with HBSS, and then incubated for 10 min at 37 C with 3 ml of freshly prepared 0.25% trypsin (Difco, Detroit, Mich.) dissolved in phosphate-buffered saline. Trypsinization results in removal of surface adsorbed toxin so that background radioactivity is reduced to a minimum and only intemalized (trypsin-insensitive) protein is measured. The cells were transferred into centrifuge tubes and spun for 2 min at low speed. The resulting pellets were washed twice in HBSS and the cells were suspended in 2 ml of cold distilled water. An equal volume of 12% trichloroacetic acid was added to each tube, and the suspensions were stored overnight at 4 C. The precipitates were collected on Mitex filters (Millipore Corp., Bedford, Mass.), washed twice with 6% trichloroacetic acid and digested in 2 ml of 0.1 N NaOH for 60 min at 56 C. Aliquots of the digests were assayed for radioactivity and protein content. Data are expressed as counts per minute per milligram of protein and represent the average of between three and six separate determinations for each point. Toxicity. Inhibition of protein synthesis was used as an index of toxicity. Cell monolayers were incubated with medium containing [3H ]leucine (Amersham-Searle Corp., Arlington Heights, Ill.) or I4C Ileucine (New England Nuclear Corp., Boston, Mass.) for indicated times, and incorporation of amino acid into trichloroacetic acid-precipitable material was determined as described elsewhere (22). Results are expressed as incorporation of [14C ]leucine or [3H Jleucine in counts per minute per microgram of protein. Polystyrene latex particle-protein preparations. Polystyrene latex spheres (1.1-Mm diameter; Dow Chemical Co., Indianapolis, Ind.) were coated with diphtheria toxin or with human serum albumin by the following method. A concentrated suspension of beads was mixed with the appropriate protein in potassium phosphate buffer (0.01 M, pH 6.9) on a tilt table for 45 min at room temperature. The suspensions were then centrifuged, and the pellets were washed twice in buffer and resuspended to the desired concentration in NCTC-135 medium plus 2% guinea pig serum. Both direct measurement of total protein and calculation of counts per minute of 125I-labeled diphtheria toxin in the pellet indicated that 4 to 5% of the original protein in the concentrated mixture was adsorbed to the latex spheres.

RESULTS Adsorption and uptake of ['251]diphtheria toxin by HEp-2 and L-929 cells. Our objective in the initial experiments was to determine whether diphtheria-susceptible and -resistant cells differ in their capacity to adsorb and/or internalize toxin. It was considered important that this be established since one published report (29) suggested that resistant mammalian cells may lack the capacity to bind toxin at the cell surface and therefore are able to withstand

the adverse effects of toxin. Because very small numbers of diphtheria toxin molecules suffice to kill a sensitive cell (18), the components of the experimental system had to be rigorously monitored to reduce the possibility of artifact. Initial experiments were performed with toxin iodinated by the chloramine T method (19). Although this procedure yields iodinated protein of high specific activity, the strong oxidizing conditions to which the proteins are subjected may result in loss of biological activity. Biologically altered preparations were discarded since the tissue and cell adsorption characteristics of toxoid and denatured toxin are quite different from those of biologically active toxin (2). The data presented in Fig. 1 provided the first indication that diphtheriaresistant L-929 cell fibroblasts do not owe their resistance to a failure to interact with toxin. L-929 cells in monolayer culture adsorbed and internalized [125I]diphtheria toxin during a 3-h period to a greater extent than did their sensitive counterparts, the HEp-2 cells. Pinocytosis of proteins by mammalian cells is temperature dependent (10, 31), and this was reflected by 9,750O

7,7501

L-CELL-37C

5,250 -

4,250_ t 3,250 HEp-2 CELL-37C k 2,250

B

1,250

HEp-2 CELL-SC

.* X L-CELL-5C 250 250

2 HOURS

3

4

FIG. 1. Adsorption and uptake of ['25Ildiphtheria toxin by HEp-2 and L-929 cells. Monolayers of HEp-2 or L-929 cells were incubated at 5 or 37 C for up to 3 h in tissue culture medium containing ['25I]diphtheria toxin (2.9 x 106 counts/min per ml, 1.1 gg/ml; chloramine T method). The level of cell-associated radioactivity was determined 30, 60, 90, and 180 min after addition of iodinated toxin. Monolayers were washed four times with HBSS, scraped loose from the plastic surface, and processed as described in Materials and Methods.

678

BONVENTRE ET AL.

the low level of radioactivity associated with both cell cultures incubated at 5 C. We interpreted the increase in cell-associated radioactivity with time as the result of pinocytosis of iodinated toxin, and this is supported by the data shown in Fig. 2. Poly-L-ornithine stimulated uptake of toxin several-fold. This was consistent with Ryser's observations (32) that high-molecular-weight polycations stimulate pinocytosis by mammalian cells after their attachment to the cell membrane at multiple sites. Adsorption and uptake of toxin in the presence of ammonium chloride. The iodination of toxin protein for the following series of experiments was carried out by the solid-state lactoperoxidase method (12), which has several advantages over the chloramine T procedure. The most important advantage is that the iodinated product sustains no alteration or measurable loss of biological activity. In addition, labeling is limited almost exclusively to the major protein component, with insignificant incorporation of iodine into lactoperoxidase; therefore the iodinated toxin is of high specific activity and purity. Another significant change in the protocol for these experiments was that the cell monolayers, after exposure to iodinated toxin, were treated with 0.25% trypsin for 10

INFECT. IMMUN.

min at 37 C rather than being scraped as in the previous experiments. In view of these modifications, uptake of toxin iodinated by lactoperoxidase bound to Sepharose beads rather than by chloramine T was reexamined. In this experiment, uptake of toxin by HEp-2 and L-929 cell monolayers was followed in the presence and absence of protective amounts (4 x 10-s M) of ammonium chloride (Fig. 3). The pattern that emerged was similar to that obtained with toxin iodinated by chloramine T oxidation (Fig. 1 and 2). As before, L-929 cells exhibited a greater extent of adsorption and subsequent uptake of toxin than did HEp-2 cells (Fig. 3). The data also show conclusively that ammonium chloride had no measurable effect on the quantity of toxin adsorbed to the cell surface or ultimately internalized. We assume that the toxin uptake measured in this assay is an index primarily of pinocytosis, although the small amounts of toxin taken into the cell by other mechanisms would also be included in the total radioactivity measured. The observations were somewhat unexpected and the data bear on two important issues. First, they show conclusively that the ammonium salt does not negate toxicity for the diphtheria-sensitive HEp-2 cell by preventing the binding of toxin to the cell membrane. This fact was also established by indirect means 2400 r 0-

ezy,

L&J zz

0.Ai

HOURS

HOURS FIG. 2. Effect of poly-L-ornithine on uptake of [1251]diphteria toxin. HEp-2 cell monolayers were exposed to ['25I]diphtheria toxin (1.9 x 106 counts! min per ml, 0.8 Ag/ml; chloramine T method) in the presence or absence of poly-L-ornithine (10 Ag/ml). Monolayers were processed as indicated in Fig. 1.

FIG. 3. Uptake of ['25I]diphtheria toxin in a medium containing or lacking ammonium chloride. [12 II]diphtheria toxin (5 x 106 counts/min per ml, 1.3 Ag/ml; labeled by solid state lactoperoxidase method) was incubated with HEp-2 or L-929 cells for 5 min before removal of the zero time sample. One set of plates of each cell type was maintained in the presence of 4 x 10-I M NH4CL during the entire 3-h incubation period. At the end of each uptake period, monolayers were washed, trypsinized, and processed as described in Materials and Methods.

described in the previous paper (22). Secondly, the data show that diphtheria-resistant cells are not immune from toxin action because of a failure to interact with toxin at the cell surface. Toxicity of iodinated toxin for HEp-2 cells. In an effort to delineate the kinetics of toxicity and uptake of toxin protein simultaneously, an experiment utilizing ["Clleucine and [125I]diphtheria toxin was performed. The use of the double label allowed assessment of protein synthesis and uptake of toxin in the same set of HEp-2 cell cultures. lodinated toxin (1.5 x 106 counts/min per ml; 2.9 mg/ml) and ["4C]leucine (0.1 uCi/ml) in tissue culture fluid were added to the HEp-2 cell monolayers simultaneously, after which protein synthesis and toxin uptake were monitored (Fig. 4). Iodinated toxin inhibited incorportion of ["4C]leucine into protein after 2 h of incubation in the absence of ammonium chloride (Fig. 4B). Incubation with the ammonium salt, however, completely negated the expression of this toxicity, although it did not alter uptake of the toxin. Addition of antitoxin to cell cultures after a 3-h incubation with

r

2600

-'2200k j900 E 700

0 C40-

*NH C

-

6500-

Z

iodinated toxin and ammonium salt prevented expression of surface-adsorbed toxin during an additional 3-h period with ammonium chloride absent. Uptake of toxin by the HEp-2 cells proceeded normally in either case (Fig. 4A). These observations provided evidence that pinocytized toxin may not express biological activity. They show also that ammonium chloride does not prevent pinocytic uptake but does act to maintain that toxin responsible for cell death in a state amenable to antitoxin neutralization. Uptake of iodinated BSA by HEp-2 and L-929 cells. It appeared pertinent to compare the adsorption and uptake of a protein other than diphtheria toxin by the cells under investigation. The rationale was to ascertain whether toxicity per se might alter the manner in which a protein would be handled by the cell or, alternatively, whether both the nontoxic and toxic substances behaved similarly in the uptake system. Serum albumin was chosen since its molecular weight and net charge are similar to that of whole diphtheria toxin. Crystalline BSA iodinated by the chloramine T method and incubated with HEp-2 cell cultures behaved essentially the same as iodinated toxin in both adsorption and uptake characteristics (data not shown). Additionally, the uptake of [121 I]BSA by HEp-2 cells did not appear to be altered significantly by toxic levels of unlabeled toxin. Uptake of BSA was not suppressed during the 3-h incubation with toxin (Fig. 5). It is interest-

ooo-

2000-

600 -

300-

L-CELLS-37C

200J

302 HOURS

679

MAMMALIAN CELLS AND ['25I]DIPHTHERIA TOXIN

VOL. 11, 1975

3

4 HOJRS

6

FIG. 4. Toxicity of iodinated diphtheria toxin for HEp-2 cells. (A) HEp-2 cell monolayers were incubated with [12 1]diphtheria toxin (1.5 x 106 counts! min per ml, 2.9 t.g/ml; lactoperoxidase method) in the presence or absence of NH4CI (4 x 10-I M). Uptake of iodinated toxin was measured as described in Fig. 3. (B) HEp-2 cell monolayers were incubated simultaneously with [125I]-diphtheria toxin (1.5 x 106 counts! min per 2.9 ug/mt) and [14C]leucine (0.1 uCi/mI) in the presence or absence of NH4CI (4 x 10-3 M). Symbols: (-) Protein synthesis in the presence of toxin alone. (A) Protein synthesis in the presence of toxin plus NH4CI. (0) Cell monolayers were maintained in the presence of iodinated toxin, [14C]leucine, and NH4CI for 3 h; at this time the monolayers were washed and reincubated for an additional 3 h in tissue culture fluid containing only antitoxin (10 U/ml) and [14C]leucine. Protein synthesis was assayed in all cases by determining incorporation of I14C]leucine into trichloroacetic acid-insoluble materials as described in Materials and Methods.

\ 1500/ 84

-

1000

/

X HEp CELLS-37C

500 5C

2

3

HOURS

FIG. 5. Effect of unlabeled diphtheria toxin on the uptake of iodinated BSA. Tissue culture medium containing ['251]BSA (2.6 x 106 counts/min per ml, 1.0 ig/ml: chloramine T method) and unlabeled diphtheria toxin were added to monolayers of HEp-2 and L-929 cells. Toxin was used at a concentration of 0.4 and 5.0 flocculating units/mI with the HEp-2 and L-929 cells, respectively. Uptake of radioactivity was determined as described in Fig. 1. Adsorption and uptake kinetics at 5 C were the same for both HEp-2 and L-929 cells.

680

INFECT. IMMUN.

BONVENTRE ET AL.

ing to note that although the level of toxin used inhibited protein synthesis in HEp-2 cells within 2 h, pinocytotic uptake of BSA did not cease during the remainder of the incubation period when little or no de novo protein synthesis was occurring. We did not attempt to preincubate HEp-2 cells with toxin for prolonged periods of time and then measure uptake of BSA. On the basis of Cohn's investigations with mouse macrophages (9), one would predict that eventually pinocytosis of protein would be halted, since metabolic inhibitors of protein synthesis were reported by him to suppress pinocytosis. The exposure of L-929 cells to diphtheria toxin was not long enough to cause inhibition of protein synthesis in these relatively resistant mouse cells (27; unpublished observation) and thus, not unexpectedly, uptake of BSA was unaffected. The data also reflect the temperature dependence of pinocyto-

NH4 CI

12Q000 8000 -0

¢

0~~~~~~~~~~~~~~~~~~~~ 4000 2 HOURS

3

FIG. 6. Effect of ammonium chloride on the uptake and toxicity of iodinated diphtheria toxin with guinea pig macrophages. The uptake of [1251]diphtheria toxin (3.4 x 108 counts/min per ml, 1.1 Ag/ml; lactoperoxidase method) by cultured guinea pig peritoneal macrophages was measured in the presence or sis. Uptake of iodinated toxin by guinea pig absence of NH4CI (4 x 10-s M). Labeled toxin was the cell monolayers; after a 5-min adsorption peritoneal macrophages. In view of the fact added tosamples were taken at 0, 1, 2, and 3 h and that phagocytic cells are innately endocytic in period, in Fig. 3. (Insert) The level indicated as processed to greater pinocytize nature and thus likely macrophages was compared by synthesis protein of quantities of protein than epithelial cells and after a 3-h exposure to iodinated toxin in the presence fibroblasts, we decided to examine the uptake of (0) or absence (U) of NH4CI (4 x 10-3 M). Protein iodinated diphtheria toxin by cultured guinea synthesis was measured by determining the incorporapig peritoneal cells. Macrophages from glyco- tion of [14C]leucine (0.1 0Ci/ml) into trichloroacetic gen-stimulated guinea pigs were cultured at a acid-insoluble materials during a 1-h pulse. concentration of 3 x 106 cells per 60-mm plastic petri dish in NCTC-135 medium and used ap- third hour a decrease in radioactivity was noted, proximately 4 h after the adherent cell popula- as in the case of the cells protected from toxin tion was established. Toxin for these experi- action by the ammonium salt. In spite of the

ments was iodinated by the lactoperoxidase method and used immediately after preparation. Uptake of [l25I]toxin (3.4 x 106 counts/ml per ml) was followed during a 3-h period with quadruplicate samples taken at hourly intervals and processed as described for the established cell lines. The uptake of toxin in the presence of protective levels of ammonium chloride proceeded at a rapid rate for 2 h and then abruptly ceased, with total radioactivity declining significantly during the third hour (Fig. 6). It should be noted that the amount of protein internalized by the macrophages during the incubation period was an order of magnitude higher than achieved by the HEp-2 cells (Fig. 2). This probably reflects the pinocytic capacity of each cell type. When the ammonium salt was excluded from the uptake mixture, however, an entirely different picture emerged. During the initial hour of incubation, toxin accumulation by the macrophages proceeded in a normal fashion but then continued at a much slower rate when compared with uptake in the presence of ammonium chloride. During the

large quantity of toxin internalized, presumably by an endocytic mechanism, no toxicity was expressed in the macrophage cultures incubated with ammonium chloride, but maximum toxicity was noted in the case of cells incubated with iodinated toxin in the absence of the ammonium salt (Fig. 6). The iodinated toxin was tested by guinea pig assay and found to possess full biological activity. On the basis of this assay, it was estimated that the macrophages were exposed to > 100 minimum lethal doses per ml, a quantity far greater than required to cause cessation of protein synthesis in 2 h (C. Saelinger, P. F. Bonventre, and J. Imhoff, J. Infect. Dis., in press). Therefore, it is apparent that the guinea pig macrophages ceased to pinocytize toxin after protein synthesis was halted, and the data suggest that the toxin, by vitrue of its unique biological activity, ultimately arrests its own pinocytosis. It also strengthens the observation with HEp-2 cells described previously (Fig. 1-4) that pinocytosis of toxin per se does not necessarily constitute a lethal event since addition of antitoxin to cells

VOL. 11, 1975

681

MAMMALIAN CELLS AND [125IJDIPHTHERIA TOXIN

after considerable toxin is internalized results in rescue of the cells. The implication of these findings is that under normal circumstances, pinocytosis of toxin may not represent the physiologically significant mode of entry of toxin into the diphtheria-sensitive cell. The fate of toxin in pinocytic vesicles probably is degradation to peptides and amino acids by lysosomal enzymes (15). This is consistent with the observation that the radioactivity during the late stages of incubation of [125I]toxin and macrophages decreased significantly, since labeled iodotyrosine diffuses out of the cell after proteolytic destruction of pinocytized protein (14). Expression of toxicity by toxin within phagocytic vacuoles. Although the previous observations suggested that toxin within phagocytic vesicles was degraded before biological activity could be expressed, we reexamined the situation by another means of toxin internalization. A mixture containing latex spheres at an optical density at 660 nm of 2.0 and purified toxin at a concentration of 4.6 mg/ml was mixed for 45 min. A control mixture using human serum albumin (0.5%) instead of toxin was treated in the same manner. After washing and centrifugation, the protein-coated latex spheres were resuspended in medium to a final optical density of 0.6. Guinea pig macrophage monolayers were then allowed to phagocytize the diluted suspensions of latex spheres in tissue culture fluid for 45 min at 37 C. Phagocytosis was monitored by phase contrast microscopy and it was estimated that each cell ingested approximately 100 latex spheres. To prevent expression of any toxicity at the level of the plasma membrane during phagocytosis, ingestion of spheres was carried out in the presence of fully protective levels of ammonium chloride (22). The monolayers were then exposed to diphtheria antitoxin (10 U/ml) to neutralize any toxin at the outer cell surface. It was reasoned that antitoxin would have no effect on toxin within phagocytic vacuoles. The fluids were then replaced with complete tissue culture medium, and protein synthesis was measured at 3, 8, and 21 h after phagocytosis by [3H ]leucine incorporation into macrophage protein during a 2-h pulse. The data shown in Fig. 7 are unequivocal. Since all toxin at the cell surface was neutralized, the toxicity expressed in this case can only be accounted for by the toxin residing within the phagosomes. If one assumes that 1% of the latex spheres were phagocytized (estimate based on level of radioactivity in macrophages after phagocytosis of iodinated toxincoated spheres) and 3 to 4% of toxin in the

NO PARTICLES

~30O 20

~~~~~TOXIN-PARTICLES

10

4

8

2

16

20

24

HOURS

FIG. 7. Expression of biological activity by toxin within phagocytic vacuoles. Cultures of guinea pig peritoneal macrophages were allowed to phagocytize polystyrene latex particles coated with diphtheria toxin or with normal human serum albumin for 45 min. Phagocytosis was carried out in the presence of NH4CI (4 x 10-I M). After phagocytosis and neutralization of toxin on the plasma membrane with antitoxin (10 U/ml), monolayers were reincubated in tissue culture medium. Incorporation of [3H]leucine (0.2 ACi/mI) into protein was determined at indicated times during a 2-h pulse.

original preparation was adsorbed to beads, then one can calculate that macrophages internalized approximately 1 to 2 minimum lethal doses per monolayer or about 1010 molecules of toxin per cell. This is an enormous quantity of intracellular toxin when compared with the estimated 2,000 molecules per cell internalized -by HEp-2 cells during a 3-h period of pinocytosis, and thus it is not surprising that some toxin would escape proteolytic digestion and express toxicity from within the cell. Effect of ammonium chloride on toxin residing within phagocytic vacuoles. In view of the fact that endocytic vesicles are membranebound structures with a reverse orientation to that of the plasma membrane, and in view of our previous observations suggesting that ammonium chloride protects sensitive cells by maintaining physiologically significant toxin at the cell surface, we examined the possibility that the salt might continue to exert its protective effect on toxin within phagocytic vacuoles. It was assumed that the limiting factor would be the diffusibility of ammonium chloride through the plasma membrane and the membranes constituting the phagocytic vacuoles. Guinea pig macrophages were allowed to

682

BONVENTRE ET AL.

phagocytize toxin-coated latex spheres as described above. The experiment was modified by maintaining ammonium chloride in the medium after the latex beads had been ingested as well as during the [3H ]leucine pulse. The results of two such experiments are shown in Table 1. When one compares these data with those shown in Fig. 7, it is clear that ammonium chloride exerted a significant influence on the expression of toxin within the phagocytic vacuole. The effect appeared to be concentration dependent, since 1.2 x 10-2 M ammonium chloride provided a higher degree of protection than the lower concentration of 4 x 10-3 M. It should be pointed out that the lesser concentration of ammonium salt provides full protection of HEp-2 cells against toxin exposure at the outer cell membrane, where diffusion would not be a limiting factor (22).

DISCUSSION Several conclusions can be drawn and a number of questions raised by these observations. In the model provided by established cell cultures of diphtheria-sensitive and -resistant lines, it appears that resistant cells do not inherently lack the capacity to adsorb toxin to the cell surface or to internalize toxin. Resistance of mouse L-929 monolayers, therefore, cannot be attributed to a failure of these cells to carry out either of these processes. Whether or not this assertion can be generalized is problematical. Pappenheimer and Brown (29) observed that only a minute amount of toxin was bound to mouse L-cell membranes. The contradictory nature of these observations might be explained by the fact that their experiments were conducted with cells grown in spinner TABLE 1. Protective action of ammonium chloride on toxin present within phagocytic vacuoles of guinea pig

macrophages NH4Cl Expt

concn post-

Albumina

Toxina

Inhibition

Noneb

28.1

4 x 10-4 Mc

29.0 35.0 35.0 26.5

7.6 15.9

45

phagocytosis

1 2

[3H Jleucine incorporation (counts/min per ug of protein)

Noneb 4 x 10-3 Mc 1.2 x 10-IMC

7.7

19.6 21.4

74 78 44 19

a Material adsorbed to latex spheres before phagocytosis. bAmmonium chloride present only during phagocytosis of protein-coated spheres. c Ammonium chloride present during phagocytosis and thereafter through the [3Hlleucine pulse.

INFECT. IMMUN.

culture and maintained in suspension during the experimental manipulations. The surface topography and configuration of mammalian cells adhering to glass or plastic in monolayer culture contrasts markedly with the same characteristics of cells suspended in a tissue culture fluid. We have compared cells in monolayer and those in suspension with respect to incorporation of [3H ]leucine into cell protein and found that the former are considerably more active (unpublished observations). This might suggest a higher total metabolic potential of cells in monolayer and thus reconcile our observations with those obtained by Pappenheimer and Brown, who used a suspended cell system (29). Our data show that both diphtheria-sensitive and -resistant cells in culture internalize significant amounts of toxin. Cellular uptake of toxin by a pinocytotic process, however, does not necessarily result in cell destruction. This was shown by the fact that ['25I]toxin internalized by HEp-2 cells during incubation with ammonium chloride does not express biological activity while the ammonium salt is present or if antitoxin is added to cells after pinocytotic uptake of toxin and the removal of the protective agent. In addition, toxin adsorbed to HEp-2 cell cultures with ammonium chloride in the medium is detoxified within 6 to 12 h, probably by cell proteases (22). Thus, since ammonium chloride does not hinder pinocytotic uptake of toxin, and specific antitoxin can rescue cells even after prolonged periods of incubation, toxin within endocytic vesicles probably does not under normal circumstrances reach the cytoplasmic constituents and express biological activity. If pinocytosis is not the physiologically important mechanism by which toxin reaches the cytoplasm of susceptible mammalian cells, then another mechanism may also be operative. One such possibility is that toxin traverses the plasma membrane without being enclosed within membrane-bound vesicles. Such a mechanism has been described for the escape of pancreatic enzymes from zymogen granules into the cytoplasm of acinar cells (25). Ribonuclease is thought to enter several plant and animal cells by a comparable mechanism (8, 17, 24). In spite of these observations, the process of direct passage of proteins through membranes is uncommon and poorly understood. Gill et al. (18) proposed several interesting models by which such transmembrane passage of whole toxin or fragment A might occur, but as yet no experimental evidence to support any of the models is available. Any model conceived must explain how a water-soluble protein penetrates the hydrophobic lipid barrier imposed by the cell membrane. The fluid mosaic model (34) of the

VOL. 11, 1975

MAMMALIAN CELLS AND [1251]DIPHTHERIA TOXIN

mammalian plasma membrane predicts that proteins may be integrated into membrane components that may move freely within the plane of the membrane. Direct passage through the membrane would require either an exorbitant expenditure of energy or a novel mechanism by which the hydrophobic lipids are circumvented. Gill (personal communication) proposes a possible mechanism for the entrance of cholera toxin into cells; the model suggests that five subunits of protein B interact with the membrane and arrange themselves in such a way as to permit protein A, the moiety responsible for cholera toxin's activation of adenyl cyclase (30), to reach the cytosol. Such a mechanism has not been seriously considered for the entrance of diphtheria toxin into mammalian cells. The major difficulty in studying the phenomenon of direct membrane transport of diphtheria toxin is to distinguish between the small number of molecules required to kill a cell (18) in comparison with the large amount of toxin that may be pinocytized. This imposes a still unresolved technical difficulty, and thus direct observations of toxin-cell interactions have not provided definitive answers to the questions of how toxin ultimately reaches its site of action in the cytoplasm. It might be appropriate to speculate how these and other observations with cultured cells pertain to the natural diphtheria infection or intoxication of susceptible species. In the guinea pig, the extent to which toxin inhibits protein synthesis is dependent upon the dose (7), the route of injection (6), and the regimen used to elicit the toxemia (3). In the acute situation where death occurs within 24 to 30 h, protein synthesis is significantly reduced in a number of tissues (2) but does not appear to be affected in others (3). In a chronic model of diphtheritic toxemia in the guinea pig, protein synthesis appears to remain relatively normal, even during the terminal stage of toxemia (4). Thus it is not clear to what extent shut-down of protein synthesis contributes to the total pathophysiology and ultimate demise of the host. In addition, studies thus far conducted with guinea pigs have not identified those proteins most adversely affected, although some evidence for direct cardiotoxicity induced by diphtheria toxin has been obtained (5). Whether or not the apparent wide range of sensitivity observed for the tissues of a host injected with toxin parenterally can be attributed to the same phenomena that confer resistance or susceptibility to L-929 and HEp-2 cells, respectively, is still unclear. The data we presented in this communication established that adsorption and pinocytic uptake occur in both cells. On the

683

other hand, we have not yet shown whether L cells lack specific toxin receptors responsible for binding a small number of toxin molecules or lack the mechanism that may allow the direct passage of this small number of molecules tLrough the membrane to reach the cytoplasm. The situation is complicated further when one attempts to assess the guinea pig findings and extrapolate to humans. The recent observation of Iglewski and Rittenberg (20) that human skin, heart, and pancreas cells cultured from adult tissues obtained at autopsy were resistant to large quantities of toxin is surprising and noteworthy. Our previous work with guinea pigs showed that heart and pancreatic tissues are quite sensitive to toxin action (4, 5), and thus it would appear that comparable tissues from the two species may respond entirely differently to toxin. On the other hand, it is not clear whether the cultures tested by Iglewski and Rittenberg were true heart and pancreatic cells; it is conceivable that fibroblast-like cells constituted a large percentage of the cultures established and this might explain the apparent resistance demonstrated. Their observation that resistance of cells disappears with the onset of malignant transformation is of great potential import (20). It reinforces the fact that toxin susceptibility or resistance is determined at the level of the plasma membrane, where attention should now be directed. Finally, since continuous cell lines, regardless of the tissues of origin, assume several abnormal characteristics usually associated with malignant transformation (16), interpretation of data regarding diphtheria toxin generated with such cell cultures should be made with discretion until comparable studies using primary cell cultures and the intact host can be conducted. ACKNOWLEDGMENTS This investigation was supported by Public Health Service research grant AM-08632 from the National Institute of Allergy and Infectious Diseases. B. I. and M. A. were supported by Public Health Service training grant GM155 from the National Institute of General Medical Sciences. LITERATURE CITED 1. Andrewes, F. W., W. Bullock, S. R. Douglas, G. Dreyer, A. D. Gardner, P. Fildes. J. C. G. Ledingham, and C. G. L. Wolf. 1923. Diphtheria, its bacteriology, pathology and immunology. Medical Research Council, Lon-

don.

2. Baseman, J. B., A. M. Pappenheimer, Jr., D. M. Gill, and A. A. Harper. 1970. Action of diphtheria toxin in

the guinea pig. J. Exp. Med. 132:1138-1152. 3. Bonventre, P. F. 1973. Studies on the mode of action of diphtheria toxin. V. Protein metabolism in a guinea pig model simutating chronic diphtheritic toxemia. Infect Immun. 7:556-560. 4. Bonventre, P. F., and J. G. Imhoff. 1966. Studies on the mode of action of diphtheria toxin. II. Protein synthein guinea pig tissues. J. Exp. Med. 124:1107-1122.

684

BONVENTRE ET AL.

INFECT. IMMUN.

5. Bonventre, P. F., and J. G. Imhoff. 1967. Studies on the mode of action of diphtheria toxin. II. Protein synthesis in primary heart cell cultures. J. Expt. Med. 126:1079-1086. 6. Boventre, P. F., and C. B. Saelinger. 1972. Inhibition of protein synthesis after intravenous or intramuscular challenge with diphtheria toxin. Infect. Immun.

mone of high radioactivity. Biochem. J. 89:114-123. 20. Iglewski, B. H., and M. B. Rittenberg. 1974. Selective

6:418-421. Bowman, C. G., and P. F. Bonventre. 1970. Studies on the mode of action of diphtheria toxin. III. Effect on subcellular components of protein synthesis from the tissues of intoxicated guinea pigs and rats. J. Exp. Med. 131:659-674. Brachet, J. 1954. Effects of ribonuclease on the metabolism of living root-tip cells. Nature (London) 174:876-877. Cohn, Z. A. 1966. The regulation of pinocytosis in mouse macrophages. I. Metabolic requirements as defined by the use of inhibitors. J. Exp. Med. 124:557-571. Cohn, Z. A., and E. Parks. 1967. The regulation of pinocytosis in mouse macrophages. II. Factors inducing vesicle formation. J. Exp. Med. 125:213-232. Collier, J. 1967. Effect of diphtheria toxin on protein synthesis: inactivation of one of the transfer factors. J. Mol. Biol. 25:83-98. David, G. S. 1972. Solid state lactoperoxidase: a highly stable enzyme for simple gently iodination of proteins. Biochem. Biophys. Res. Commun. 48:464-471. Duncan, J. L., and N. B. Groman. 1969. Activity of diphtheria toxin. II. Early events in the intoxication of HeLa cells. J. Bacteriol. 98:963-969. Ehrenreich, B. A., and Z. A. Cohn. 1967. The uptake and digestion of iodinated human serum albumin by macrophages in vitro. J. Exp. Med. 126:941-963. Ehrenreich, B. A., and Z. A. Cohn. 1969. The fate of peptides pinocytosed by macrophages in vitro. J. Exp. Med. 129:227-245. Fenner, F., B. R. McAuslan, C. A. Mims, J. Sambrook, and D. 0. White. 1974. The biology of animal viruses. Academic Press Inc., New York. Firket, H., S. Chevremont-Comhaire, and M. Chevremont. 1955. Action of ribonuclease on living cells in vitro and synthesis of deoxyribonucleic acid. Nature (London) 176:1075-1076. Gill, D. M., A. M. Pappenheimer, Jr., and T. Uchida. 1973. Diphtheria toxin, protein synthesis and the cell. Fed. Proc. 32:1508-1515. Greenwood, F. C., W. M. Hunter, and J. S. Glover. 1963. The preparation of 126I-labelled human growth hor-

22. Ivins, B., H. F. Bonventre, C. B. Saelinger, and C. Woscinski. Chemical modulation of diphtheria toxin

7.

8.

9. 10. 11. 12.

13. 14.

15.

16. 17.

18. 19.

toxicity of diphtheria toxin for malignant cells. Proc.

Nat. Acad. Sci. U.S.A. 71:2707-2710. 21. Ittelson, T. R., and D. M. Gill. 1973. Diphtheria toxin: specific competition for cell receptors. Nature (Lon-

don) 242:330-332.

action

on

cultured mammalian cells. Infect. Immun.

11:000-000.

23. Johnson, W., R. J. Kuchler, and M. Solotorovsky. 1968. Site in cell-free protein synthesis sensitive to diphtheria toxin. J. Bacteriol. 96:1089-1098. 24. Ledoux, L. 1955. Action of ribonuclease on certain ascites tumours. Nature (London) 175:258-259.

25. Liebow, C., and S. Rothman. 1972. Membrane transport of proteins. Nature (London) New Biol. 240:176-178. 26. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall. 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193:265-275. 27. Moehring, J. M., and T. J. Moehring. 1968. The response of cultured mammalian cells to diphtheria toxin. II. The resistant cell: enhancement of toxin action by

poly-L-ornithine. J. Exp. Med. 127:541-554. 28. Oyama, V. I., and H. Eagle. 1956. Measurement of cell growth in tissue culture with a phenol reagent (FolinCiocalteau). Proc. Soc. Exp. Biol. Med. 91:305-307. 29. Pappenheimer, A. M., Jr., and R. Brown. 1968. Studies on the mode of action of diphtheria toxin. IV. Site of action of toxin in living cells. J. Exp. Med. 127:1073-1086. 30. Pierce, N. F., N. B. Greenough III, and C. C. J. Carpenter. 1971. Vibrio cholerae enterotoxin and its mode of action. Bacteriol. Rev. 35:1-13. 31. Ryser, H. J. P. 1963. The measurement of '3'I-serum albumin uptake by tumor cells in tissue culture. Lab.

Invest. 12:1009-1017. 32. Ryser, H. J. P. 1967. A membrane effect of basic polymers dependent on molecular size. Nature (London) 215:934-936. 33. Ryser, H. J. P. 1968. The uptake of foreign proteins by mammalian cells and the functions of pinocytosis. Bull. Swiss Acad. Med. Sci. 24:363-385. 34. Singer, S. J., and GX. L. Nicolson. 1972. The fluid mosaic model of the structure of cell membranes. Science 175:720-731.

Interaction of cultured mammalian cells with [125I] diphtheria toxin.

The characteristics of cell adsorption and pinocytotic uptake of diphtheria toxin by several mammalian cell types were studied. Purified toxin iodinat...
2MB Sizes 0 Downloads 0 Views