Effect of lactoferrin on enteroaggregative E. coli (EAEC).

We previously demonstrated that lactoferrin inhibits adherence of enteropathogenic Escherichia coli to HEp-2 cells and decreases invasiveness of Shigella flexneri in HeLa cells by disruption of the type III secretory system (TTSS) of both enteropathogens. To determine whether these effects were specific to the TTSS, we assessed the activity of bovine lactoferrin on enteroaggregative E. coli (EAEC), enteropathogens whose virulence is not TTSS dependent. Bovine lactoferrin at a concentration of 1.0 and 0.1 mg/mL inhibited EAEC growth. Saturation with iron reversed the bacteriostatic effect. Lactoferrin under nonbacteriostatic conditions decreased EAEC adherence to HEp-2 cells as evaluated by microscopy and CFUs; this effect was not iron dependent. Lactoferrin inhibited EAEC biofilm formation and increased autoagglutination. Lactoferrin blocks EAEC adherence by inducing release and degradation of aggregative adherence fimbria, a key element of EAEC pathogenesis. We hypothesized that lactoferrin binding to lipid A of lipopolysaccharide disrupts the virulence proteins anchored to the bacterial outermembrane. These data suggest that the effect of lactoferrin on surface proteins is not restricted to organisms having a TTSS.     

Macrophage binding of cells resistant and sensitive to contact-dependent cytotoxicity

We compared macrophage binding and killing of F5b cells to the binding and killing of P815 mastocytoma cells and to several other nontransformed and transformed cell lines. Formalin fixation of elicited or activated macrophages did not affect binding of F5b or 3T3 cells but did abrogate binding of P815 cells. However, formalin fixation abrogated resident macrophage binding of F5b and 3T3 cells. Therefore, depending on the type of macrophage or target cell, formalin fixation may affect binding. Only the binding of P815 cells was dependent upon activation; macrophage binding of target cells F5b and 3T3 was not. Even though macrophages bound F5b and 3T3 cells, macrophages only mediated contact-dependent cytotoxicity against F5b cells. Macrophages did not kill 3T3 cells. Experiments also compared macrophage binding and killing of the uv-light-induced tumor cell lines 1422, 2237, and 2237a46. Only the cell line 2237a46 was susceptible to contact-dependent killing. Both 1422 and 2237 cells were resistant. In contrast, cell lines 2237a46 and 1422 were bound by activated macrophages while 2237 cells were bound poorly.     

The range of oxygenation in SiHa tumor xenografts

Tumor cells at very low oxygen tensions are known to be about three times more resistant to killing by ionizing radiation. Since cells at intermediate oxygen tensions (defined here as greater than 0.1% and less than 2% O(2)) show partial radioresistance, they should be a consideration in tumor treatment. In an effort to estimate the extent and range of oxygenation in SiHa human cervical carcinoma xenografts, patterns of cell killing and DNA damage by radiation and two bioreductive drugs, PD-144872 and RSU-1069, were compared to those seen in SiHa cells grown as spheroids. These drugs produce DNA interstrand crosslinks that are largely responsible for cell killing, and the degree of crosslinking increases as the oxygenation is reduced. About 60% of the cells in SiHa xenografts exhibited drug-induced crosslinks, but only about 35% showed extensive crosslinking indicative of hypoxia below 0.1% oxygen. Patterns of toxicity and DNA damage in xenografts were comparable to those of spheroids equilibrated with about 2% oxygen, indicating that most cells in the xenografts exhibit some radioresistance due to lack of oxygen. Similarly, pimonidazole binding indicated that about 60% of the cells in SiHa xenografts were either intermediate in oxygenation or hypoxic, but only about half of those were consistent with extreme oxygen depletion. The apparent size of the population of "intermediately hypoxic" cells has implications for the use of ionizing radiation, hypoxic cell cytotoxins, and other antitumor agents whose cytotoxicity is dependent on cellular oxygen content.     

Splenocytes cultured in low concentrations of IL-2 generate NK cell specificities toward syngenic and allogenic targets

Splenocytes cultured in the presence of 30-60 units/ml IL-2 for 5 days develop natural killer activity toward syngeneic and allogeneic tumor cell targets. The IL-2 activated splenocytes, themselves, are partially resistant, whereas concanavalin A-activated T blast cells are completely resistant to killing. Surprisingly, major histocompatibility complex (MHC)-I-negative target cells are also resistant to natural killer (NK)-cell-mediated killing. Cells resistant to killing were unable to block NK-cell-mediated killing of sensitive targets as judged from cold target cell inhibition experiments, and one type of target cells sensitive to killing did generally not cross-block killing of other killing-sensitive target cell types. Alloantigen exposure of splenocytes, i.e., one-way mixed lymphocyte cultures, partially prevents the development of NK-cell activity. Our data suggest that target structures which trigger killing activity of NK cells are determined by the phenotype of the target cell and are dependent on its MHC class I expression disregarding the haplotype of the cell.     

Ferric iron and superoxide ions are required for the killing of cultured hepatocytes by hydrogen peroxide. Evidence for the participation of hydroxyl radicals formed by an iron-catalyzed Haber-Weiss reaction

Cultured hepatocytes pretreated with the ferric iron chelator deferoxamine were resistant to the toxicity of H2O2 generated by either glucose oxidase or by the metabolism of menadione (2-methyl-1,4-naphthoquinone). Ferric, ferrous, or cupric ions restored the sensitivity of the cells to H2O2. Deferoxamine added to hepatocytes previously treated with this chelator prevented the restoration of cell killing by only ferric iron. The free radical scavengers mannitol, thiourea, benzoate, and 4-methylmercapto-2-oxobutyrate protected either native cells exposed to H2O2 or pretreated hepatocytes exposed to H2O2 and given ferric or ferrous iron. Superoxide dismutase prevented the killing of native hepatocytes by either glucose oxidase or menadione. With deferoxamine-pretreated hepatocytes, superoxide dismutase prevented the cell killing dependent upon the addition of ferric but not ferrous iron. Catalase prevented the killing by menadione of deferoxamine-pretreated hepatocytes given either ferric or ferrous iron. Deferoxamine pretreatment did not prevent the toxicity of t-butyl hydroperoxide but did, however, prevent that of cumene hydroperoxide. It is concluded that both ferric iron and superoxide ions are required for the killing of cultured hepatocytes by H2O2. The toxicity of H2O2 is also dependent upon its reaction with ferrous iron to form hydroxyl radicals by the Fenton reaction. The ferrous iron needed for this reaction is formed by the reduction of cellular ferric iron by superoxide ions. Such a sequence corresponds to the so-called iron-catalyzed Haber-Weiss reaction, and the present report documents its participation in the killing of intact hepatocytes by H2O2. Cumene hydroperoxide but not t-butyl hydroperoxide closely models the toxicity of hydrogen peroxide.     

Serum resistance in bvg-regulated mutants of Bordetella pertussis

Serum resistance, or resistance to killing by antibody dependent pathway of complement, in Bordetella pertussis is bvg-regulated and the Bordetella resistance to killing (brk) locus mediates much of the resistance. Here we examined whether other bvg-regulated proteins contribute to serum resistance. We found that neither pertussis toxin, adenylate cyclase toxin, filamentous hemagglutinin, dermonecrotic toxin, tracheal colonization factor, nor Vag8 mutants were sensitive to serum killing compared to the wild-type. Filamentous hemagglutinin has been reported to bind C4 binding protein, an inhibitor of complement, but this activity does not appear to contribute to serum resistance, as evidenced by the resistant phenotype of FHA mutants. Clinical isolates were serum resistant and wild-type strains possessing an additional copy of the brk locus were 2–5-fold more resistant to serum killing.     

Binding of yeast killer toxin to a cell wall receptor on sensitive Saccharomyces cerevisiae.

35S-labeled killer toxin protein bound to cells of sensitive Saccharomyces cerevisiae S14a. Strains that were resistant to toxin through mutation in the nuclear genes kre1 kre2 bound toxin only weakly. Non-radioactive toxin competed effectively with 35S-labeled toxin for binding to S14a, but did not compete significantly in the binding to mutant kre1-1. This implied that binding to kre1-1 was nonspecific. A Scatchard analysis of the specific binding to S14a gave a linear plot, with an association constant of 2.9 x 10(6) M-1 and a receptor number of 1.1 x 10(7) per cell. Killer toxin receptors were solubilized from the cell wall by zymolyase digestion. Soluble, non-dialyzable cell wall digest from S14a competed with sensitive yeast cells for 35S-labeled toxin binding and reduced toxin-dependent killing of a sensitive strain. Wall digest from kre1-1 competed only weakly for toxin binding with sensitive cells and caused little reduction of toxin-dependent killing. Although the abundant (1.1 x 10(7) per cell) receptor appeared necessary for toxin action, as few as 2.8 x 10(4) toxin molecules were necessary to kill a sensitive cell of S14a. The kinetics killing of S14a suggested that some component was saturated with toxin at a concentration 50-fold lower than that needed to saturate the wall receptor.     

The effect of interferon on lymphocyte-mediated effector cell functions: selective enhancement of natural killer cells.

The effect of human interferons on different types of lymphocyte-mediated killer assays was explored. Killing by T cells generated through mixed lymphocyte cultures as well as antibody-dependent lymphocyte-mediated cytotoxicity was not influenced by the addition of interferon. Enhancement of cytolysis produced by natural killer cells was observed when interferon was added during the assay, but enhancement could also be induced if the effector cells were pretreated with interferon for 2 hr prior to the lytic reaction. Killing of a cell line susceptible to natural killing was increased and a cell line which is normally relatively resistant to this type of killing became a susceptible target.     

The effect of CaCl2 and verapamil on the binding of cisplatin to cultures of normal and transformed human fibroblasts

Normal and transformed human fibroblasts were treated for either 1 sec or 1 h with the antitumor drug cis-dichlorodiamine platinum (cisplatin). The dose response of drug binding and cell survival was determined for cells treated with the drug in the presence or absence of 3.0 mM CaCl2. The levels of drug initially bound to both cell types was similar and was not affected by the presence of Ca2+. The dividing non-transformed cells were most sensitive to killing by short treatment with cisplatin compared to the transformed cells or the confluent non-transformed cultures. After 1 h of cisplatin treatment, the levels of drug bound to the cells were significantly less than that recovered after the shorter treatment. This time-dependent loss of cisplatin was inhibited both by CaCl2 and by the calcium channel blocking agent, verapamil. The higher levels of cisplatin bound after 1 h in the presence of these agents, however, did not in all cases result in decreased survival; the effects were dependent on cell type and on whether the cells were dividing or confluent. Analysis of cisplatin binding to cell cultures indicated that initially the cisplatin was weakly attached to the pericellular and substratum attached material but that with time, the drug bound to this material decreased. This time-dependent removal from the extracellular matrix was much less in the transformed cell cultures and was inhibited by calcium. We propose that the major site of interaction of cisplatin with these cells is in the extracellular matrix and with time the cultures alter their extracellular matrix to decrease this binding. This removal process appears to involve calcium or calcium transport since CaCl2 and verapamil both block these changes.     

The effect of monensin on the antibody-dependent cell-mediated killing

Antibody-dependent cell-mediated immune cytolysis of herpes simplex virus-infected target cells involves two separate events, recognition and adhesion of the effector K cells to the antibody-coated target cells, and the final killing step. In the present study the killing event is shown to be dependent on an intact secretory mechanism in the K cells. Treatment of K cells with the carboxylic ionophore monensin, which blocks secretion, completely abolishes the K-cell-mediated killing, but the adhesion of the effector cells to the antibody-coated target cells is not affected by monensin. The similarity between the killing events mediated by K and NK cells is discussed.     

Augmentation by anti-T3 antibody of the lymphokine-activated killer cell-mediated cytotoxicity

This study showed that a mAb (145-2C11) against the T3 epsilon-chain of the TCR complex augmented the cytotoxic activity of the lymphokine-activated killer (LAK) effectors. The LAK cells were induced by culturing normal spleen cells with purified human rIL-2. Adding alpha T3 at the effector phase of the cytotoxic reactions augmented the LAK-mediated cytotoxicity. The alpha T3-augmented LAK killing was seen only with tumor targets, and there was no increase of killing against Con A-induced lymphoblasts. The augmentation effect was dose dependent on both the amounts of alpha T3 and the number of LAK cells added. A very low concentration of alpha T3 (1/10,000 dilution of culture supernatants) was sufficient to induce alpha T3-augmented LAK-mediated cytotoxicity. Human rIL-2 at 10 to 30 U/ml was sufficient to generate LAK cells for maximal alpha T3 augmentation, whereas 300 to 1000 U/ml of IL-2 were needed to generate maximal LAK activity when tested in the absence of alpha T3. LAK cells generated for longer periods of time showed a progressive increase of alpha T3-augmented cytotoxicity. For some targets, the alpha T3-augmented LAK killing was FcR dependent as evidenced by the ability of alpha FcR mAb 2.4G2 to inhibit, and for others it was not inhibited. The alpha T3-augmented killing did not correlate with the FcR expression on target cells as defined by 2.4G2. The LAK cells were both Lyt-2+ and Lyt-2-, but the LAK cells involved in alpha T3-augmented killing were exclusively Lyt-2+. Preincubation of LAK cells with alpha T3, but not preincubation of targets with alpha T3, resulted in augmented killing suggesting that the alpha T3 effect was unrelated to an antibody-dependent cell-mediated cytotoxicity. Our findings indicate that alpha T3 is a potent reagent to augment the cytotoxic reaction of LAK cells. These results suggested that a relationship might exist between the T3 complex and the cytotoxic activity of a subpopulation of Lyt-2+ LAK cells.     

The effect of temperature or anoxia on Escherichia coli killing induced by hydrogen peroxide

The cytotoxicity of hydrogen peroxide in Escherichia coli was investigated after various conditions of drug exposure. Two modes of killing were detected following a 15-min challenge with H2O2 under either aerated or anoxic conditions. Mode one killing occurred at levels below 2.5 mM and mode two killing at concentrations higher than 10 mM. Whereas mode one killing was similar at the two conditions of drug exposure, mode two lethality differed in that aerated cells were more sensitive than anoxic cells. Independently of O2 tension the hydroxyl radical scavenger, thiourea, prevented mode two but not mode one killing by H2O2. Cells treated with the drug at ice temperature did not display mode one killing and mode two lethality occurred only at very high concentrations. We suggest that hydroxyl radicals mediate mode two but not mode one killing by H2O2.     

Potassium channels in human NK cells are involved in discrete stages of the killing process

Using the whole-cell variation of the patch-clamp technique, we have found a voltage-dependent K+ current in human natural killer (NK) cells. This K+ current is reduced in a dose-dependent manner by a variety of ion-channel blockers (verapamil, quinidine, 4-aminopyridine, Cd2+) at concentrations comparable to those that inhibit natural killing. Pretreatment of target cells with quinidine or verapamil did not significantly reduce their sensitivity to killing, whereas substantial inhibition of killing was observed after pretreatment of effector cells. Both verapamil and quinidine reduced the proportion of effector-target cell conjugates, suggesting that K channels play a role in the "binding" phase of the killing process. By adding EDTA or channel blockers as various times in a Ca-pulse assay system, we have also delineated a blocker-sensitive phase of bound conjugates that strictly corresponds with the Ca-dependent "programming" stage of killing. In contrast, the killer cell-independent stage, which is Ca2+ independent, apparently does not require functioning K channels. Verapamil and quinidine do not affect target cell sensitivity to the putative soluble mediator of natural killing, natural killer cytotoxic factor (NKCF), but inhibit release of NKCF from NK cells. Thus, the data suggest that K channels in NK cells play essential roles in the natural killing process that include events in the "programming-for-lysis" phase leading to release of NKCF.     

The effect of 45 degrees C hyperthermia on the membrane fluidity of cells of several lines.

The membrane fluidity of cells of human (AG1522 human foreskin fibroblasts), rodent [Chinese hamster ovary (CHO) and radiation-induced mouse fibrosarcoma], and feline (Crandall feline kidney) cell lines after heating at 45 degrees C was measured by flow cytometry. In addition, a heat-resistant variant of radiation-induced mouse fibrosarcoma cells and two heat-sensitive CHO strains were studied. Fluorescence polarization of the plasma membrane probe trimethylammonium-diphenylhexatriene was used as a measure of membrane fluidity. The sensitivity of all cell lines to 45 degrees C hyperthermia was compared. The baseline membrane fluidity varied among the cell lines, but did not correlate with sensitivity to hyperthermia. However, CHO cells, especially the heat-sensitive mutants, had the largest increase in membrane fluidity after heating at 45 degrees C, while the heat-resistant mouse fibrosarcoma variants and Crandall feline kidney cells resisted changes in fluidity. In general, the more resistant the cell line was to killing by heat, the more resistant it was to changes in membrane fluidity.     

A complement-resistant HeLa cell line (T638) is blocked at the step of C3 deposition

A complement-resistant line of HeLa cells (T638) was derived by serial passage of complement-susceptible HeLa cells in anti-beta 2-microglobulin (b2m) antiserum and complement. The T638 line maintained stable complement resistance when passed for an additional 1500 generations in the absence of antiserum and complement. T638 cells expressed equivalent levels of cell-associated b2m as did the parent HeLa cell line. Furthermore, T638 cells were resistant to killing by complement and anti-HeLa antiserum with specificity for molecules other than b2m. These results indicate that the resistance of T638 cells does not simply reflect loss of anti-b2m binding antigens. We next investigated the mechanism of resistance of T638 cells to complement-mediated killing. Antibody-sensitized HeLa and T638 cells both consumed CH50 activity completely from normal human serum; cytotoxicity was not mediated via the alternative complement pathway. HeLa and T638 cells caused equivalent utilization of C4 from normal human serum in the presence of antibody. Consumption of C2, greater with T638 than with HeLa cells during incubation in serum, was complete when cells bearing purified C1 and limited C4 were incubated with C2. T638 cells bound more 3H-C4 than HeLa cells during incubation in serum, but binding of 3H-C3 by T638 cells was fourfold to fivefold less than by HeLa cells. Finally, we investigated the rate of decay in the capacity of C142 on HeLa and T638 to cleave and deposit 3H-C3. The T1/2 for decay of C142-mediated binding of 3H-C3 on HeLa was 3.9 min, whereas minimal C3 deposition was detected on T638 cells at all time points. These results show that T638 cells evade complement-mediated lysis despite activating early components of the classical complement pathway. The mechanism of resistance is a failure to form an effective C3 convertase.     

Lymphocyte-mediated lysis of sheep chorion: susceptibility of chorionic cells to third-party and maternal cytotoxic lymphocytes and presence of cells in the endometrium exhibiting cytotoxicity toward natural-killer cell targets

In several species, the trophoblast is resistant to lysis by cytotoxic lymphocytes. Such resistance is believed to contribute to survival of the semiallogenic conceptus. We tested whether ovine chorionic cells are susceptible to lysis by specific and nonspecific cytotoxic lymphocytes in peripheral blood (PBL) and whether cytotoxic cells that can lyse target cells for natural-killer cells are present in the endometrium. Primary chorionic cells from pregnant ewes at Days 51-91 of gestation were labeled with 51Cr and incubated for 20 h at 50:1 and 100:1 ratios with PBL from the pregnant mother or from a third-party ewe. In the absence of interleukin-2 (IL-2), there was no killing of primary chorionic cells by third-party PBL even after infection of chorionic cells with bovine herpes virus-1. Incubation with IL-2-induced cytotoxic action in third-party PBL towards one of six primary chorionic cell preparations only. Primary chorionic cells from two of four placentae were lysed by maternal PBL. Luminal epithelial cells from cyclic ewes and from the pregnant and nonpregnant uterine horns of unilaterally-pregnant ewes were evaluated for the presence of cells capable of killing D17 target cells (a natural-killer cell target). Killing was observed but there was no difference in activity between physiological stages. In contrast, there was intense immunochemical localization of perforin in glandular and luminal endometrial epithelial cells in pregnant ewes, and less intense staining in nonpregnant animals. It is concluded that ovine chorionic cells are generally resistant to killing by natural-killer-like cells and lymphokine-activated killer cells. Generation of maternal cytotoxic lymphocytes against trophoblast can occur in some cases and may contribute to pregnancy loss.     

Multiple, small exposures of far-ultraviolet or mid-ultraviolet light change the sensitivity to acute ultraviolet exposures measured by cell lethality and mutagenesis in V79 Chinese hamster cells

Chinese hamster V79 cells (subline MI2G) were exposed repeatedly to fractionated doses of germicidal 254 nm light (far-uv) at 6 J.m-2/fraction/day or sunlight-simulating 290-330 nm (mid-uv) at 150 J.m-2/fraction/day and sensitivities to cell killing action and mutation of far-uv and mid-uv were examined. As the number of exposure fractions increased, the cell cultures became resistant to cell killing induced by both far-uv and mid-uv. Increases in both Do and Dq were observed. Treatment with exposures of 6 J.m-2 far-uv is more efficient in yielding cell cultures that are resistant than exposures of 150 J.m-2 mid-uv. In contrast to the cells exposed to repeated far-uv, the cells exposed to repeated mid-uv were relatively more resistant to cell killing effects of mid-uv than far-uv, suggesting a possible role of photolesions other than pyrimidine dimers. When mutants resistant to 6-thioguanine were assayed during repeated exposure to far- or mid-uv light, the yield was initially linear with accumulating dose. At high total accumulated doses, the frequency decreased gradually (6 J.m-2 mid-uv) or reached a plateau (150 J.m-2 mid-uv). The sensitivity of N80 cells (exposed to 80 fractions of mid-uv) to mutation induction by uv light is higher than that of the original MI2G cells, whereas U81 cells (exposed to 81 fractions of far-uv) have a sensitivity similar to that of the original cells. Although an initial decrease in resistance to cell killing was observed, resistant cells retained their characteristics after 100 days in culture without further exposure. Cross-resistance to X rays was not shown. The data in this paper suggest that the capacity for repair of photolesions in DNA by repair processes was enhanced in cell cultures by repeated exposure to far-uv or mid-uv and that this altered the cells' ability to cope with lethal and mutagenic lesions. It remains to be seen if these changes in cell sensitivity were brought about by selective or inductive processes or a combination of both.     

Radiosensitivity of human natural killer cells: binding and cytotoxic activities of natural killer cell subsets

The sensitivity of human natural killer (NK) cell activities (both binding and killing) after exposure of peripheral blood mononuclear cells to different doses of gamma radiation was studied. A panel of monoclonal antibodies was used to identify the NK and T-lymphocyte subsets and to evaluate their radiosensitivity. Peripheral blood mononuclear cells were irradiated with low (2-6 Gy) and high (10-30 Gy) doses and NK cell binding and cytotoxic activity against K562 target cells were studied after 3 h and 48 h in culture. The primary damage to NK cell activity was identified at the postbinding level and affected mainly the lytic machinery. After 48 h culture postirradiation, an overall depression of cytotoxic activity was observed, but ionizing radiation produced either a selection of the more cytotoxic NK cell subsets, which therefore might be considered more resistant to radiation damage than the less cytotoxic NK cells, or a long-term stimulation of cytotoxic activity in surviving cells.     

Characterization of macrophage recognition and killing of SV40-transformed tumor cells that are "resistant" or "susceptible" to contact-mediated killing

In this report we used the macrophage-"resistant" and -"susceptible" cell lines, F5m and F5b, to determine why AKR or AKR-like virus expression makes the F5m cell line more resistant to in vitro macrophage killing than the F5b cell line. We found that resistance to macrophage killing may be transmitted by an infectious AKR or AKR-like murine leukemia virus and that resistance was concomitant with virus expression as measured by the presence of AKR virus-specific 70 kDa glycoprotein. We report that macrophage cytotoxicity of these cell lines is dependent upon the direct contact between tumor cells and macrophages. In contrast, macrophage-mediated cytostasis occurred via soluble macrophage products and no differential susceptibility of F5b or F5m to macrophage-mediated cytostasis was observed. Macrophage binding of F5b was also significantly better than that of F5m. These data suggest that only the events that depend upon the close contact of macrophages and tumor cells will be affected by the expression of AKR or AKR-like virus. Therefore, the differences in susceptibility of F5m and F5b to direct macrophage-mediated cytotoxicity are apparently because the macrophage binding of F5m is less efficient than the binding of F5b.     

A molecular mechanism of complement resistance of human melanoma cells

The susceptibility of human melanoma cells to lysis by human complement after sensitization with the R24 murine IgG3 monoclonal antibody to the GD3 ganglioside antigen was investigated. It was found that the melanoma cell lines were either susceptible (greater than or equal to 70% cytotoxicity) or resistant (less than or equal to 30% cytotoxicity) to complement-mediated killing. We determined the kinetics of binding of C3 to and its subsequent fate on the melanoma cells. We found that on susceptible cell lines, maximal binding of C3 occurred within 10 min of incubation. At that time, approximately 90% of the bound C3 was in the form of C3b. During the subsequent incubation, the C3b was slowly inactivated, apparently generating the physiologic degradation products iC3b, C3dg, and C3d. However, this degradation of C3b could be inhibited without affecting the final degree of cytotoxicity, indicating that it is of no apparent consequence for the killing of susceptible melanoma cells. Very different results were obtained with resistant melanoma cells. Bound C3b was rapidly inactivated, and C3d was the predominant form of C3 on resistant cells throughout the incubation. Therefore, rapid inactivation of C3b was identified as a protective mechanism of human melanoma cells against complement attack. In addition, we found that resistance to complement is not an inherent property of the cells but depends on the antibody used for sensitization, because the resistant cell lines could be lysed after sensitization with polyclonal antiserum.