[CANCER RESEARCH 53. 592-599. February I. 1993]

Molecular Basis of Complement Resistance of Human Melanoma Cells Expressingthe C3-cleaving Membrane Protease p65l

Markus W. Ollert,2 Joseph V. Kadlec, Eugene C. Petrella, Reinhard Bredehorst, and Carl-Wilhelm Vogel

Departments of Biochemistry and Molecular Biology. Medicine, and Pediatrics, and the tnternalitmal Center for Interdisciplinary Studies of Immunology, Georgetown UniversitySchool of Medicine, Washington. DC 20007 ¡M.W. O.. J. V. K.. E. C. P.. R. B.. C-W. V.l. and the Department of Biochemistry and Molecular Biologv. University of Hamburg.Hainburg, Germany IM. W. O.. K. B.. C-W. VJ

ABSTRACT

The molecular mechanism of complement resistance of the humanSK-MKL-170 melanoma cell line was investigated. The cells have beenshown to express the C3b-cleaving membrane protease p65. To delineatethe molecular consequences of the C3b-cleaving activity for the comple

ment cylotoxicity, the molecular events during the initiation (R24 monoclonal antibody. CD. amplification (C4, C3), and membrane attack (C5,C9) phases of complement were studied in comparison to a complement-susceptible human melanoma line (SK-MEL-93-2). No cleavage of C4b

and (51). 2 molecules structurally similar to C3b, was observed on the cellsduring classical pathway activation indicating the specificity of the p65protease for the C3b molecule. The rapid degradation of C3b by p65 on thesurface of complement-resistant SK-MKL-170 cells generates a U, 30,000C3(«'-chain-fragment detectable as early as 1 min after complement acti

vation, whereas no such fragment was present in detectable amounts oncomplement-susceptible cells. As a result of the rapid C3b proteolysis byp65 on resistant SK-MKL-170 cells, less C5 convertases are formed, which

in turn results in the formation of a lower number of terminal complementcomponents and membrane attack complexes. R24 antibody and Clqbinding to the resistant cells was slightly lower as to susceptible cells. C4binding studies, however, revealed that the observed difference in antibody and Clq binding has no influence on the complement resistance ofSK-MKL-170 cells: significantly more C4b was bound to complement-

resistant (1565 ±92 fg/cell) as compared to susceptible cells (715 ±31fg/cell). On extraction of the molecular forms of C4 bound to the cellmembranes, an additional high molecular weight C4 species—apparentlya C4b-C4b homodimer—appeared only on the resistant SK-MKL-170cells that may function as a residual back-up C5 convertase. Collectively,these results show that SK-MKL-170 human melanoma cells evade complement-mediated cytolysis despite sufficient activation of early components of the classical complement pathway by p65-mediated rapid degradation of surface-bound C3b, leading to a significant reduction in

membrane attack complex formation. Thus, rapid cleavage of surfacedeposited C3b was established as a powerful mechanism of complementresistance.

INTRODUCTION

The complement system is an efficient mediator of antibody-induced tumor cell cytolysis (3). Thus, complement-activating niAb'

have been widely studied to inhibit various human tumors (6-8).

Many of the antitumor effects of mAb against human tumors are

Received 6/17/92; accepted 11/11/92.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby marked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

1This work was supported by NIH Grants CA35525 and CAOI039. Preliminary

accounts were presented at the 13th International Complement Workshop. San Diego. CA.September 1989 ( 1) and at the 15th International Cancer Congress, Hamburg. Germany.August 1990 (21. M. W. O. was supported at Georgetown University by postdoctoralfellowship No. 0156/1-2 from the Deutsche Forschungsgemeinschaft.

2 To whom reprint requests should be addressed, at Department of Biochemistry andMolecular Biology. University of Hamburg. Martin-Luther-King-Pl. 6. 20(X)Hamburg 13.

Germany.' The abbreviations used are: GDPBS. Dulbecco's phosphate-buffered saline ( 137 HIM

NaCl, 2.7 mi«KCI, 6.5 mu Na-phosphate, 1.5 nui K-phosphate, pH 7.4) with 0.1% (w/v)gelatin; GDPBS34. Dulbecco's phosphate-buffered saline (137 nui NaCl. 2.7 mMKCI. 6.5

nui Na-phosphale. 1.5 nut K-phosphate. pH 7.4) with 0.1% (w/v) gelatin with 0.5 mwMgCli and 0.9 imi CaCli; mAb, monoclonal antibody; MAC. membrane attack complex;NHS, normal human serum. The nomenclature of the complement components follows therecommendations of the World Health Organization (4. 5).

attributed to activation of the classical complement pathway withformation of the MAC (6). Among the best studied antitumor antibodies is the complement-activating anti-GD1 mAb R24 of the murineIgG3 subclass (3, 6, 9-11). Clinical trials have revealed regression of

metastatic melanoma concomitant with deposition of complementcomponents at tumor sites including late components of the MAC (6,12. 13). However, tumor cells vary in their susceptibility to complement-mediated killing due to several mechanisms. Some tumors are

resistant to the complement attack by expressing DAF or MCP (14,15). Another mechanism of complement resistance of tumor cells isthe expression of p65, a C3b-cleaving membrane protease, on human

melanoma cells as recently identified in our laboratory (16. 17). Therapid proteolytic degradation of C3b on the target cell surface by p65after complement activation by the R24 mAb results in the generationof a M, 30,000 C3 breakdown product on complement-resistant mel

anoma cells. Preincubation of the resistant melanoma cells with across-reacting antiserum to a C3b-cleaving erythrocyte protease ( p57)

increased the extent of complement killing and the amount of intactC3b during the activation phase of the classical pathway (17, 18). Todelineate the molecular consequences of the C3b-cleaving activity for

the complement cytotoxicity, we studied the molecular events duringthe initiation (mAb. Cl ), amplification (C4, C3), and membrane attack(C5, C9) phases of complement activation in comparison to a complement-susceptible melanoma cell line (SK-MEL-93-2).

MATERIALS AND METHODS

Melanoma Cell Lines. SK-MEL-170 and SK-MEL-93-2 human mela

noma cells were kept in continuous culture by using RPMI 1640 containing0.3% (w/v) t-glutamine (Mediatech. Washington, DC), supplemented with\0r/c (v/v) heat-inactivated fetal bovine serum (Inovar Biologicals, Gaithers-burg. MD). \c/c (v/v) nonessential amino acids (Mediatech). 1% (v/v) sodium

pyruvate (Mediatech), and I7c (v/v) penicillin-streptomycin (GIBCO Laboratories. Grand Island. NY). Cells were harvested either with trypsin-EDTA

(GIBCO) or by mechanical means. The 2 human melanoma cell lines used inthe present study have been described and characterized previously (9, 16. 17).They differ in their susceptibility to complement-mediated killing by the R24mAb using trypan blue dye exclusion (SK-MEL-170 <45% cytotoxicity;SK-MEL-93-2 >95% cytotoxicity) (16).

Antibodies and Complement Components. Ascites containing the anti-

Go, mAb R24, obtained from BALB/c mice with i.p. growing hybridomas.was used to sensitize cells in complement activation assays (9-11). For anti

body binding studies, the R24 antibody was purified through proteinG-Sepharose affinity chromatography. NHS. obtained from healthy volunteers

by cubital vein puncture, served as the source of complement in all experiments and was stored at -90°C until used. Human C3 and C5 were purified

from fresh frozen plasma (American Red Cross Blood Service. Washington.DC) as described (19). Human complement components Clq. C4. and C9 wereobtained from Quidell-Cytotech (San Diego. CA). Protein concentrations ofthe purified complement components were determined by measuring the ab-sorbance at 280 nm using extinction coefficients (in ml •¿�mg~' •¿�cm"') of e =

0.68 for Clq (20), €¿�= 1.00 for C4 (21 ), €¿�= 0.97 for C3 (22), e = 1.09 forC5 (23), and e = 0.96 for C9 (24).

Binding of I25l-labeled Complement Components to Melanoma Cells.Cells (8 X IO6) were sensitized with the R24 antibody by incubation in 3.2 mlR24 ascites (diluted 1/5 in GDPBS:\ pH 7.4) for 45 min at 0°C.Control cellswere incubated with GDPBS2 ' under identical conditions. Subsequently, the

592

Research. on August 20, 2019. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from

MOLECULAR BASIS OF COMPLEMENT RESISTANCE

cells were washed once and incubated at 37°Cin 4 ml NHS (diluted 1/2 inGDPBS2 +) spiked with the 12il-labeled complement component to be inves

tigated. The amount of 125l-labeled component added to the serum was 1/120

of the amount in normal serum except for Clq, where 1/40 was added. At thetime intervals indicated, quadruplicate aliquots containing 200.000 cells in atotal volume of 100 ul were removed, layered carefully onto cold 30% (w/v)sucrose in a 500-ul Eppendorf tube, and spun immediately for 1.5 min in an

Eppendorf microfuge. After cenlrifugation. the tubes were rapidly frozen inliquid nitrogen, the tips containing the cell pellets were cut off with a razorblade, and the radioactivity of each tip was counted in a gamma-counter. The

counts sedimented with unsensitized control cells were subtracted from theexperimental values. Nonspecific binding was in the range between 5 and 10%of the specific binding. The calculations of the amounts of each complementcomponent bound per cell are based on the respective specific activities and thenormal serum concentration of each component. The following serum concentrations were used for calculations: 0.06 mg/ml for Clq, 0.60 mg/ml for C4.1.30 mg/ml for C3,0.07 mg/ml for C5. and 0.06 mg/ml for C9 (25). Statisticalanalysis of the obtained data was performed according to the Student's t test.

Binding of '"¡-labeled R24 mAb to Melanoma Cells. Melanoma cells(2 x IO5) were incubated on ice for 45 min with various concentrations of R24mAb spiked with 1% (w/w) '"I-labeled R24 mAb in a total volume of 100 ulGDPBS2 +. The cells were then layered carefully onto 30% (w/v) sucrose and

were spun and processed as described above. The total amount of R24 boundto the cells was calculated based on the number of cpm and the specific activity

as described (26).Extraction of '"¡-labeled Complement Components from Melanoma

Cells. The extraction of complement components from the membranes ofmelanoma cells was performed as described previously for C3 (16) with slightmodifications. Briefly, melanoma cells (2.5 x IO6) were incubated with 1 mlof R24 ascites (diluted 1/5 in GDPBS2*) for 45 min at 0°C.Subsequently, thecells were washed and incubated with 2.5 ml NHS (diluted 1/2 in GDPBS2' )

spiked with the respective I2sl-labeled complement component (see above) forthe times indicated. The cells were washed in Dulbecco's phosphate-buffered

saline2* (137 min NaCI, 2.7 mm KC1, 6.5 msi Na-phosphate. 1.5 m\i K-phos-

phate, pH 7.4) containing 25 mM /)-nitrophenylguanidino benzoate (Sigma

Chemical Co., St. Louis, MO), lysed in hypotonie buffer ( I mw NaHCO,. 2 mmCaCli. 25 HIMp-nitrophenylguanidino benzoate. pH 8.0) at 0°C,the nuclei

were removed by centrifugation at 300 x g, and I25l-labeled complement

components were extracted from the plasma membranes with 1% (w/v) SDSand 25 mvi methylamine at pH II as described (16). Samples were analyzed bySDS-PAGE and subsequent autoradiography at -90°C.

Other Methods. SDS-PAGE was performed according to established pro

cedures (27) in 7.5% (w/v) and 9% (w/v) minigels. For the analysis of membrane-bound C9. a 2.5-10% (w/v) gradient gel was overlaid with a 2.5% (w/v)

separating gel and a 2.5% (w/v) stacking gel as described (24). Radiolabelingof the R24 mAb and the complement components Clq, C4, C3, C5, and C9was performed with Na'2il (Amersham) by using immobilized chloramine-T(Iodo Beads: Pierce, Rockford, IL). Nonincorporated 12ÕIwas separated from

the labeled proteins by gel filtration through Sephadex G25 (Pharmacia, Inc..Piscataway, NJ). Stock solutions had specific activities in the range of 2.9 to3.8 X IO6 cpm/ug (R24 mAb). 3.13 to 4.40 x 10" cpm/ug (Clq), 1.83 to 2.94X IO6 cpm/ug (C4), 5.87 to 8.75 x IO5 cpm/ug (C3), 2.44 to 3.18 x 10"cpm/ug (C5), and 3.49 to 5.30 x 10" cpm/ug (C9).

RESULTS

Binding of C3. Maximum binding of C3 to melanoma cells wassignificantly higher on complement-susceptible SK-MEL-93-2 cells(4894 ±29 fg/cell) as compared to complement-resistant SK-MEL-

170 cells (1125 ±143 fg/cell) (Table 1; Fig. 1). To gain informationon the kinetics of C3 degradation and the size of the resulting degradation products, we extracted C3 at various time points from resistantand susceptible melanoma cells. As expected, on susceptible melanoma cells the majority of the a'-chain of C3b is degraded to the Mr68,000 a'-chain fragment of iC3b within 1.5 to 3 min of incubation.

In addition, small amounts of 2 low molecular weight species of Mr45,000 and 30,000. respectively, were detectable (Fig. 2). After 10 minof incubation, iC3b is the predominant form of C3 on susceptible

Table 1 Minimum amounts ofClq, C4, C3, C5, and C9 on complement-resistant and-susceplible human melanoma cells

Maximum amount/cell (fg)"

Cell line Clq C4 C3 C5

SK-MEL-170 53±7 1565±92 1123 ±143 71 ±11 81 ±15SK-MEL-93-2 88 t 19 715 ±31 4894 ±29 283 ±58 207 ±2

" Mean ±SD of 4 independent experiments.

5000

INCUBATION AT 37 C [min]

Fig. 1. Kinetics of binding of C3 to susceptible (SK-MEL-93-2, O) and resistant(SK-MEL-170, •¿�)human melanoma cells. Shown is I of 4 representative experiments.

cells. This form is present on the cell surface for at least 30 min. Incontrast, on complement-resistant melanoma cells a M, 30.000a'-chain-derived breakdown product is detectable on the cells as earlyas 1 min after incubation. This Mr 30,000 a'-chain fragment is linked

via a disulfide bond to the intact ß-chain(Mr 72,000), as SDS-PAGE

under nonreducing conditions reveals a C3 species with an apparentmolecular weight of 102,000 (data not shown). This degradation product consisting of the M, 30.000 a'-chain fragment and the ß-chain

becomes the predominant form of C3 within 5 min and the onlydetectable form after 15 min of incubation on the resistant melanomacells (Fig. 3). In addition to the M, 30.000 fragment, small amounts ofM, 45,000 and 25,000 fragments are detectable on resistant cells onlywithin the first 10 min of incubation (Fig. 3).

Binding of C4. Maximum binding of C4 is reached after 5 min ofincubation on both melanoma cell lines. Surprisingly, the complement-resistant cells (1565 ± 92 fg/cell) bound more than twiceas much C4 than complement-susceptible cells (715 ±31 fg/cell)

(Fig. 4; Table 1). As expected, C4b is converted to C4d (M, 47.000)on complement-susceptible cells, whereas the C4b was found to be

resistant to degradation on the resistant cells even after prolongedincubation (60 min) (Fig. 5). In addition, a methylamine-resistant M,

\ 70.000 C4 species was detectable under reducing conditions after 5min of incubation on resistant melanoma cells only (Fig. 5).

Binding of R24 Antibody and Clq. To answer the question ofwhether complement resistance of SK-MEL-170 cells was a conse

quence of a significant lower binding of antibody and Clq, we determined the binding of R24 mAb and Clq on SK-MEL-170 andSK-MEL-93-2 cells. On binding studies with increasing amounts ofI25l-labeled R24 over a concentration range of 0.2 ug/ml up to 100

(jg/ml saturation was not observed on either cell line (Fig. 6); rather,the known homophilic binding properties of the R24 antibody causeda continuous increase in binding. At 200 ug/ml. saturation seemed tohave been reached on the SK-MEL-93-2 cells only (Fig. 6). At this

concentration (the approximate concentration used for sensitization incomplement binding studies), which represents a 10-fold excess of the

593

Research. on August 20, 2019. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from

MOL.KCUl.AR BASIS OF COMP1.KMKNT RESISTANCE

a

ß

—¿�105

Fig. 2. Forms of C3 extracted from susceptibleSK-MEL-93-2 cells. The cells were sensitized andincubated in normal human serum containing I2?l-

labelcd C3. After time periods as indicated, extractsof cells were prepared and subjected to SDS-PAGEunder reducing conditions with subsequent autorad-

iography. Apparent molecular masses (in kilodallons)are shown.

C3b

—¿�45

—¿�30

0.5 1 1.5 10 15

INCUBATION TIME (min)

approximate number of antigens expressed on both cell lines, the totalamounts of R24 mAb bound to the cells differed by a factor of only1.6 (I470 ± HO fg/cell on SK-MEL-93-2; 910 ±70 fg/cell onSK-MEL-170).

Subsequently, both melanoma cell lines were compared for theircapacity to bind Clq. Binding of Clq was very rapid on both cell linesand peaked after 10 to 20 s (88 ±19 fg/cell on SK-MEL-93-2; 53 ±7 fg/cell on SK-MEL-170). The relative difference of 1.7 in Clq

binding between the 2 cell lines closely resembled that found withR24 binding (Table I; Fig. 7). It should be noted that some of the Clqwas already bound at 0°Cbefore the incubation temperature wasraised to 37°C.Extraction of i:5I-labeled Clq from the membranes of

lysed melanoma cells revealed a single band of M, 25.000 underreducing conditions on both cell lines, which represents the knownsize of the Clq subunits (Fig. 7, inset).

Binding of Terminal Complement Components. Subsequently,the effect of the rapid C3b degradation on resistant cells on the extentof binding of the terminal complement components was studied. Asshown in Fig. 8, the binding kinetics of C5 on both cell lines resem

bled those obtained with C3: maximum binding of C5 was reachedwithin 10 min of incubation on both cell lines, with the complement-susceptible SK-MEL-93-2 cells binding approximately 4 times asmuch C5 (283 ±58 fg/cell) as the complement-resistant SK-MEL-

170 cells (71 ±II fg/cell) (Table I). During subsequent incubation,the amount of C5 detectable on both cell lines decreased with time.The extraction of 125I-labeled C5 from both cell lines showed no

apparent breakdown of C5b on either line (Fig. 9).The molecular forms of C9 deposited on resistant and susceptible

melanoma cells are shown in Fig. K). Monomeric, dimeric. and circularly polymeric C9 was detected on both cell lines at each timepoint. As judged from the signal intensity, the total amount of C9deposited onto susceptible SK-MEL-93-2 cells was significantlyhigher as onto the resistant SK-MEL-170 cells. Quantitation by bind

ing experiments revealed that the maximum amount of C9 depositedon the susceptible cells (207 ±2 fg/cell) was 2.6-fold higher than on

the resistant cells (81 ±15 fg/cell) (Fig. 11: Table 1). As with C5, acontinuous decrease of the amount of C9 with time was observed withboth cell lines.

Fig. 3. Forms of C3 extracted from resistant SK-MEL-170 cells. The cells were sensitized and incubated in normal human serum containing '-^-labeled

C3. After time periods as indicated, extracts of cellswere prepared and subjected to SDS-PAGE underreducing conditions with subsequent autoradiogra-phy. Apparent molecular masses (in kiltxialtons) areshown.

C3b 0.5

—¿�30

1.5 10 15

INCUBATION TIME (min)

Research. on August 20, 2019. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from

MOl.l-rUl.AR BASIS OK COMPLEMENT RESISTANCE

1600-

1200-

800-

400-

uou.o

!INCUBATION AT 37°C [min]

Fig. 4. Kinetics of binding of C4 to susceptible (SK-MEL-93-2. O) and resistant(SK-MEL-170. •¿�)human melanoma cells. Shown is I of 4 representative experiments.

DISCUSSION

We have previously shown that complement-resistant SK-MEL-170melanoma cells express a C3b-cleaving membrane protease (p65)

(17). The present study was designed to delineate the specificity andmolecular consequences of p65 activity and therefore to identify themolecular basis of the complement resistance of SK-MEL-170 human

melanoma cells.The present results indicate that the protease p65 is highly specific

for the C3 molecule. Both C4 and C5, although structurally verysimilar to C3 (28), are not cleaved by p65. Within the C3 molecule themembrane protease p65 leads to the generation of a A/r 30,000 a'-

chain fragment bound to the intact ß-chain.Since the Mr 30,000a'-chain fragment must contain the thioester region of C3, these

results indicate that one putative cleavage site of pf>5 is located between the 2 cysteines 794 and 851 of the NH2-terminal part of the C3a-chain, while a second putative cleavage site may reside within the

a

ß—mm —¿�75Fig. 5. Fonns of C4 extracted from susceptible

SK-MEL-93-2 cells and resistant SK-MEL-170 cells.

The cells were sensitised and incubated in normalhuman serum containing l:sl-laheled C4. After time

periods as indicateli, extracts of cells were preparedand subjected to SDS-PAGE under reducing condi

tions with subsequent autoradiography. Apparent molecular masses (in kilodaltons) are shown.

C3d part of the C3 a-chain (Fig. 12). The results would furthersuggest that Cys 794 is involved in C3 a-chain intrachain-, and Cys851 in interchain-disulfide bonding of the C3 molecule.

The data of this study demonstrate that the specific cleavage of C3bby the protease p65 is the principal molecular event responsible for thecomplement resistance of SK-MEL-170 cells. Upon classical pathwayactivation on the complement-susceptible SK-MEL-93-2 cells, large

numbers of C3b molecules are bound that are relatively quickly inactivated by conversion to iC3b. This relatively fast conversion of C3bto iC3b is not surprising since these cells are not activators of thealternative pathway, thereby providing basically no protection for C3bmolecules against inactivation by the serum regulatory protein factorsH and I (29). Nevertheless, despite the relatively short half-life of C3bon the surface of these complement-susceptible cells, the number of

active C3b molecules must be sufficiently high to allow for the formation of the necessary numbers of C5 convertases required forefficient complement cytotoxicity (>95% in the case of the SK-MEL-93-2 cells). It appears, therefore, that the classical pathway on a

nonactivator surface of the alternative pathway is only functional dueto massive deposition of C3b. which in turn leaves only a short kineticwindow for C5 convertase formation and C5 activation before C3binactivation occurs. In contrast, on complement-resistant SK-MEL-170 cells, the already short-lived C3b molecule is very rapidly de

graded by the p65 protease, causing the formation of a lower numberof C5 convertases, which in turn leads to less activation of terminalcomplement components and increased cell survival. Therefore, limiting the number of C3b molecules, the apparent function of the p65protease, is a powerful mechanism of protection against complementkilling. Consistent with this conclusion is our previous observationthat preincubating the resistant melanoma cells with an antiserumcross-reacting with pfi5 increases both the number of intact C3b

molecules and cell survival (17).It is interesting to note that another complement-resistant cell line

(HeLa T638) similarly showed a significantly reduced C3b bindingcompared to the complement-susceptible wild-type cells (30). While

the molecular mechanism of reduced C3b deposition on T638 HeLa

I *

Y—mm —¿�32€¿�»<•»

—¿�170

—¿�84

—¿�75

—¿�47

—¿�32

C4 5 min 30 min 60 min

SK-MEL-170

5 min 30 min

SK-MEL-93-2

595

Research. on August 20, 2019. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from

MOLECULAR BASIS OF COMPLEMENT RESISTANCE

E

I

1500

1000

500

10

R24 Added (ng'ml)

100

Fig. 6. Binding of R24 mAb to susceptible (SK-MEL-93-2, O) and resistant (SK-MEL-I70. •¿�)human melanoma cells. Shown is 1 of 2 representative experiments.

LUO

ou.O

oi

INCUBATION AT 37°C[min]

Fig. 7. Kinetics of binding of Clq to susceptible (SK-MEL-93-2, O) and resistant(SK-MEL-170. •¿�)human melanoma cells. Shown is I of 4 representative experiments.Insel, forms of Clq extracted from resistant SK-MEL-170 cells (Lane I ) and susceptibleSK-MEL-93-2 cells (Lane 2). The cells were sensitized and incubated in normal humanserum containing i:^I-labeled Clq. After 45 s of incubation, extracts of cells were

prepared and subjected to SDS-PAGE under reducing conditions with subsequent auto-

radiography. Apparent molecular masses (in kilodaltons) are shown.

cells is not known, the data reported could be fully explained by thepresence of a C3b cleaving protease. Whatever the mechanism maybe, these data corroborate our findings that limiting the amount of C3bis a sensible mechanism for complement resistance.

The molecular events taking place before the activation of C3 arenot a critical factor in the complement resistance of SK-MEL-170

cells. Welt et al. (9) showed that the average number of Gw moleculesper cell as measured by R24 antibody binding is in good correlationwith a melanoma cell line's susceptibility to complement killing.

Therefore, major differences in R24 binding between the 2 cell linesthat could be responsible for the observed differences in C3b bindinghad to be ruled out. R24 binding to the cells tested over a concentration range of 0.2 to 200 ug/ml revealed a 1.4-1.8-fold higher antibodybinding to susceptible SK-MEL-93-2 cells (Fig. 6). On the level ofClq, a 1.7-fold higher binding was found for susceptible as compared

to resistant melanoma cells (Fig 7; Table 1). The observation of an

unusually high molar ratio of bound R24 mAb to bound Clq ofapproximately 50:1 on both melanoma cell lines can be explained bythe homophilic binding properties of the R24 mAb that prevent saturation of antibody binding (26). Thus, R24 binding to melanomacells increases as the concentration of R24 increases, leaving onlycertain R24 molecules available for Cl binding. The relatively smalldifference in R24 and Clq binding between the susceptible and theresistant melanoma cell lines is unlikely to have any significant consequence on complement-mediated cytotoxicity of these cells. Further

support for this conclusion are the molecular events observed on thelevel of C4: the 1.7-fold higher binding of Clq to susceptible SK-MEL-93-2 cells should also give rise to a somewhat higher number of

C4b molecules on these cells. Quite to the contrary, the susceptibleSK-MEL-93-2 cells bind less than half the amount of C4 than theresistant SK-MEL-170 cells (Fig. 4; Table 1). The higher binding ofC4 by resistant SK-MEL-170 cells was also found with complement-

resistant T638 HeLa cells (30).Another remarkable difference was that C4b was degraded to the

physiological degradation product C4d on susceptible SK-MEL-93-2

cells from early time points on, whereas it was resistant to degradationon resistant SK-MEL-170 cells for up to 60 min of incubation (Fig. 5).

These results indicate a protective mechanism against C4b cleavageby the physiological regulatory proteins C4b binding protein andFactor I on resistant melanoma cells. The reduced C4b inactivationmight be a consequence of the low numbers of C3b and C5b on theresistant cells since a negative feedback regulation of the C4b,2aconvertase by C3b and C5b has been shown (31).

The higher amount of C4b on resistant cells not only may be causedby the reduced inactivation but may also indicate the presence ofadditional acceptor sites for C4b. Some evidence for the latter isderived from the presence of a Mr 170,000 C4 species in extracts fromresistant cells after SDS-PAGE under reducing conditions, presumably representing a C4 a'-chain dimer (Fig. 5). Such C4 a'-chain

dimer found exclusively on complement-resistant cells must be part ofa C4b-C4b homodimer. It was resistant to treatment with methy-lamine, thus ruling out the possibility of a C4b-C3b dimer that wascharacterized as having an ester-mediated binding of C3b to the a'-

chain of C4b (32). Therefore, the nature of the covalent interactionwithin the C4b-C4b homodimer seems to be an amide bond as it wasdescribed recently (33). It was found that such a C4b-C4b homodimer

can serve as a subunit of a novel C5 convertase that involves no C3derivatives (33). It was suggested that the physiological role of sucha C5 convertase (C4b,2a,4b) was to allow for residual complement

250

200

O

40

INCUBATION AT 37°C[min]

Fig. 8. Kinetics of binding of C5 to susceptible (SK-MEL-93-2. O) and resistant(SK-MEL-170. •¿�)human melanoma cells. Shown is 1 of 4 representative experiments.

596

Research. on August 20, 2019. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from

MOI.I-XVLAR BASIS OF COMPLEMENT RESISTANCE

a

Fig. 9. Forms of C5 extracted from susceptibleSK-MEL-93-2 cells and resistant SK-MEL-170 cells.

The cells were sensitized and incubated in normalhuman scrum containing I25l-labeled C5. After time

periods as indicated, extracts of cells were preparedand subjected to SDS-PAGE under reducing conditions with subsequent autoradiography. Apparent molecular masses (in kilodaltons) are shown.

—¿�116

—¿�80

—¿�105

—¿�80

C5 5 min 10min

SK-MEL-93-2

5 min 10min

SK-MEL-170

SK-MEL-93-2 SK-MEL-170

I I I I

poly-C9 —¿�

dtmericC9 —¿�

monomeric C9 —¿�

any significant effect of other lysis restricting mechanisms such as thepresence of other membrane-bound regulatory proteins (CRI. DAF,

MCP. CD59, HRF) and the active shedding and imcrnalization ofterminal complement components has to be excluded. The presence ofCRl on either melanoma cell line was ruled out directly in immuno-

precipitation assays with a polyclonal antiserum against CRl (17).Although DAF was detectable in small amounts on both cell lines,preincubation of the resistant melanoma cells with a polyclonal anti-DAF serum had no effect on complement-mediated cytotoxicity as itwas shown for melanoma cell lines that are characteri/.ed by DAF-

mediated complement resistance (14. 17). The presence of MCP onthe resistant cell line was not ruled out directly, but the known potential of MCP to act as a cofactor not only in Factor I-mediated

cleavage of C3b but also of C4b (34) makes an involvement of MCPunlikely, for no C4b cleavage occurred on SK-MEL-170 cells. Thecontribution of CD59 and HRF (C8-binding-protein) to the complement resistance of SK-MEL-170 cells is unlikely because of the

known interaction of these restriction factors with the MAC. whichresults in an inhibition of poly-C9 formation, a molecular form of C9

that was present in a similar molar ratio to monomeric C9 on bothmelanoma cell lines (Fig. 10) (35. 36).

Another important mechanism of complement resistance is theremoval of C5b-9 complexes from the membranes of nucleated cells

C9 10 10

INCUBATION TIME (min)Fig. 10. Forms of C9 extracted from susceptible SK-MEL-93-2 cells and resistant

SK-MEL-170 cells. The cells were sensitized and incubated in normal human serumcontaining '-M-labeled C9. After time peruxis as indicated, extracts of cells were preparedand subjected to 2.5-10% SDS-PAGE under reducing conditions with subsequent auto-radiography. A C9 control is shown in the left lane.

activation in C3-deficient individuals. Our observations suggest an

other possible function of such a C5 convertase: in the case of/705-bearing cells, in which rapid C3b degradation leads to ineffi

cient formation of the normal classical pathway C5 convertase(C4b,2a,3b), the formation of a backup classical pathway C5 convertase (C4b,2a,4b) may account for some residual degree of C5activation.

To allow the conclusion that the specific C3b cleavage by pf>5 is theprincipal mechanism of complement resistance on SK-MEL-170 cells.

250

200

150

100-

50.

10 20 30

INCUBATION AT 37°C[min]

40

Fig. II. Kinetics of binding of C9 to susceptible (SK-MEL-93-2. O) and resistant(SK-MEL-170. •¿�)human melanoma cells. Shown is 1 of 4 representative experiments

597

Research. on August 20, 2019. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from

MOLECULAR BASIS OF COMPLEMENT RESISTANCE

Pf5 0C ISKÃŽ*794

fMIp8

1HOOC—¿�|I2.

3.||—¿�COOHa-chain

4.5.1

NH? 6.

ß-chain

Fig. 12. Schematic representation of the C3 degradation product generated by the p65protease. The degradation product (shaded) consists of a M, 30,000 a'-chain fragment

linked to the ß-chain.

either by shedding or by a process of internalization (3, 37-39). On

both melanoma cell lines the amounts of C5b and C9 found at latertime points (30-40 min) were lower as compared to early time points(5-10 min) (compare Figs. 8 and 11). Thus, some degree of C5b-9

removal seems to occur on both cell lines. However, the absoluteamount of C9 removed from the cell surface was higher on susceptibleSK-MEL-93-2 cells (-100 fg/cell) compared to resistant SK-MEL-170 cells (—70 fg/cell). Further experimental evidence that the removal of C5b-9 is not a critical factor in complement resistance ofSK-MEL-170 cells is derived from the kinetics of cytotoxicity that

approximate maximum killing after IO min of incubation when thebinding of C5b and C9 also is near maximum on both cell lines (datanot shown). The decline of cell surface-bound C5b and C9 is observed

only at later time points, thus making it rather unlikely that theremoval of MACs is causing the complement resistance of SK-MEL-

170 cells. In conclusion, the rapid degradation of C3b by the p65protease appears to be the principal mechanism of complement resistance of human SK-MEL-170 melanoma cells.

The physiological role of the p65 protease remains to be investigated. At present, it is unknown which normal cells and tissues express the p65 protease. However, its apparent function of restrictingcomplement-dependent lysis suggests that the p65 protease is another

membrane regulatory protein that protects against accidental lysis ofhost cells by host complement along with DAF, MCP, CRI, CD59,and HRF. However, it cannot be ruled out that the p65 protease mayserve additional purposes such as generating adhesion-promotingbivalent C3 ligands on the cell surface, as it was suggested for pro-

teolytic C3 fragments on the promonocytic U937 cells and on stimulated peripheral blood lymphocytes (40, 41).

In addition to any physiological role of p65 in complement regulation as outlined above, p65 may be of clinical importance in humanmelanoma. Expression of p65 by human melanoma cells in vivo mayrepresent the molecular basis for the rather infrequent remissionsobserved in phase I clinical studies with the potent complement-

activating mAb R24 (12, 13). To address this issue, we are currentlypurifying p65 with the aim to generate /?65-specific monoclonal antibodies and -cDNA clones as reagents to screen for the expression of

pf>5in human melanoma. These results can be expected to delineate apossible clinical significance of p65 expression.

ACKNOWLEDGMENTS

The authors are indebted to Dr. Lloyd J. Old of the Memorial Sloan-

Kettering Cancer Center in New York for generously providing the R24 mAbused in these studies.

REFERENCESI. Ollert. M.. Kadlec, J. V.. Pelrella, E. C.. Bredehorst R., and Vogel, C-W. Molecular

basis of complement resistance: kinetics of binding and degradation of C3. C4 and C5on susceptible and resistant human melanoma cells. Complement Inflammation, 6:3S1-382, 1989.

10.

14.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

Ollert, M., Kadlec, J. V., Petrella, E. C., Bredehorst, R., and Vogel. C-W. Molecularbasis of complement resistance of human melanoma cells. J. Cancer Res. Clin.Oncol.. ;/6(Suppl. I): 368. 1990.Vogel, C-W. Complement, a biologic effector mechanism for tumor cell killing. In: H.

F. Oettgen (ed.). Immunology and Allergy Clinics of North America. Human CancerImmunology II, Vol. II, pp. 277-299. Philadelphia: W. B. Saunders, 1991.

Nomenclature Committee of the World Health Organization. Nomenclature of complement. Bull. WHO, 39: 935-938, 1968.

Nomenclature Committee of the World Health Organization. Nomenclature of thealternative activating pathway of complement. Bull. WHO. 59: 489^191, 1981.Chapman. P. B., Scheinberg, D. A., DiMaggio. J. J.. and Houghton. A. Unconjugaledmonoclonal antibodies as anticancer agents. In: H. F. Oettgen (ed.). Immunology andAllergy Clinics of North America. Human Cancer Immunology II, Vol. II. pp.257-275. Philadelphia: W. B. Saunders, 1991.

Capone, P. M.. Papsidero, L. D.. Croghan, G. A., and Chu, T. M. Experimentaltunioricidal effects of monoclonal antibody against solid breast tumors. Proc. Nati.Acad. Sci. USA. 80: 7328-7332, 1983.Cheung, N-K., Lazarus, H., Miraldi, F. D., Abramowsky, C. R., Kallick. S., Saaringen,

U. M., Spitzer, T., Strandjord, S. E., Coccia, P. F.. and Berger, N. A. Ganglioside Grx,specific monoclonal antibody 3F8: a phase I study in patients with melanoma andneuroblastoma. J. Clin. Oncol., 5: 1430-1440, 1987.Welt. S., Carswell, E. A., Vogel, C-W., Oettgen, H. F.. and Old, L. J. Immune andnonimmune effector functions of IgG, mouse monoclonal antibody R24 detecting thedisialoganglioside GM on the surface of melanoma cells. Clin. Immunol. Immuno-palhol., 45: 214-229, 1987.

Dippold, W. G., Lloyd. K. O., Li, L. T. C.. Ikeda. H.. Oettgen, H. F., and Old, L. J.Cell surface antigens of human malignant melanoma: definition of six antigenicsystems with mouse monoclonal antibodies. Proc. Nati. Acad. Sci. USA, 77: 6114-6118, 1980.Pukel, C. S., Lloyd, K. O.. Travassos. L. R., Dippold, W. G., Oettgen, H. F.. and Old,L. J. GD.Õ,a prominent ganglioside of human melanoma: detection and characterization by mouse monoclonal antibody. J. Exp. Med.. 155: 1133-1147, 1982.Houghton, A. N.. Mintzer, D., Cordon-Cardo, C.. Welt, S., Fliegel. B., Vadhan, S.,

Carswell, E., Melamed, M. R., Oettgen, H. F., and Old, L. J. Mouse monoclonal IgG <antibody detecting GDI ganglioside: a phase I trial in patients with malignant melanoma. Proc. Nati. Acad. Sci. USA. 82: 1242-1246, 1985.Vadhan-Raj, S., Cordon-Cardo, C.. Carswell, E., Mintzer, D., Dantis, L., Duteau, C.,

Templeton. M. A.. Oettgen, H. F., Old, L. J., and Houghton, A. N. Phase I trial ofmouse monoclonal antibody against G^ ganglioside in patients with melanoma:induction of inflammatory responses at tumor sites. J. Clin. Oncol., 6: 1636-1648,1988.Cheung, N-K. V, Walter, E. I., Smith-Mensah. W. H.. Ratnoff. W. D., Tykocinski, M.L.. and Medof, M. E. Decay-accelerating factor protects human tumor cells fromcomplement- mediated cytotoxicity in vilro. J. Clin. Invest., 81: 1122-1128. 1988.Seya, T., Hará,T.. Matsumoto, M., Sugita, Y., and Akedo, H. Complement-mediated

tumor cell damage induced by antibodies against membrane cofactor protein (MCP.CD46). J. Exp. Med., 172: 1673-1677, 1990.Panneerselvam, M.. Welt. S.. Old. L. J.. and Vogel, C-W. A molecular mechanism ofcomplement resistance of human melanoma cells. J. Immunol., 136: 2534-2541,

1986.Ollert, M., Frade, R.. Fiandino. A.. Panneerselvam, M.. Petrella, E. C.. Barel, M.,Pangbum. M. K.. Bredehorst. R.. and Vogel, C-W. C3-cleaving membrane proteinase.A new complement regulatory protein of human melanoma cells. J. Immunol.. 144:3862-3867, 1990.Charriaut-Marlangue. C., Barel. M., and Frade. R. Identification of p-57, a serineproteinase. from human erythrocyte membranes, which cleaves both chains of humanthird component (C3) of complement. Biochem. Biophys. Res. Commun., 140: 1113-

1120, 1986.Hammer, C. H., Wirtz, G. H., Renfer. L.. Gresham, H. D.. and Tack, B. F. Large scaleisolation of functionally active components of the human complement system. J. Biol.Chem.. 256: 3995-4006, 1981.Reid, K. B. M. Preparation of human Clq, a subcomponent of the first component ofthe classical pathway of complement. In: L. Lorand (ed.). Methods in Enzymology,Proteolytic Enzymes, Part C, Vol. 80, pp. 16-25. New York: Academic Press, 1981.

Isenman. D. E., and Young, J. R. The molecular basis for the difference in immunehemolysis activity of the Chido and Rogers isotypes of human complement component C4. J. Immunol., 132: 3019-3027.Tack, B. F., and Prahl, J. W. Third component of human complement: purificationfrom plasma and physicochemical characterization. Biochemistry, 15: 4513—4521,

1976.DiScipio, R. G., Smith, C. A., Müller-Eberhard. H. J., and Mugli, T. E. The activationof human complement component C5 by a fluid phase C5 convertase. J. Biol. Chem.,258: 10629-10636. 1983.Podack. E. R.. Tschopp, J.. and Müller-Eberhard. H. J. Molecular organization of C9

within the membrane attack complex of complement. Induction of circular C9 polymerization by the C5b-8 assembly. J. Exp. Med., 756: 268-282, 1982.

Reid K. B. M. The complement system. In: B. D. Hames and D. M. Glover (eds.),Molecular Immunology, pp. 189-241. Oxford: IRL Press, 1988.

Chapman, P. B., Yuasa. H., and Houghton. A. N. Homophilic binding of mousemonoclonal antibodies against GD1 ganglioside. J. Immunol., 145: 891-898, 1990.Laemmli, U. K. Cleavage of structural proteins during the assembly of the head ofbacteriophage T4. Nature (Lond.), 227: 680-685. 1970.

De Bruijn. M. H. L.. and Fey. G. H. Human complement component C3: cDNAcoding sequence and derived primary structure. Proc. Nati. Acad. Sci. USA, 82:708-712, 1985.

598

Research. on August 20, 2019. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from

MOI LfULAR BASIS OF COMPLEMENT RESISTANCE

Newman, S. L., and Mikus. L. K. Deposition of C3b and iC3b onto paniculateactivators of the human complement system: quantitation with monoclonal antibodieslo human C3. J. Exp. Med., 161: 1414-1431. 1985.Siimi. P., Joiner, K. A., Hammer, C. H.. Frank, M. M.. and Tosi. R. A complement-

ivsistant HeLa cell line (T638I is blocked at the step of C3 deposition. J. Immunol../.(.S: 3385-3391. 1987..Sirunk. R. C., and Giclas. P. C. Modulation of the activity of the classical complementpathway C3 convelíaseby surface-bound C3 or C5.J. Immunol.. 124: 520-526. 1980.

Takata, Y.. Kinoshita. T., Kozono, H., Takeda, J.. E. Tanaka. E.. Hong, K., and Inoue.K. Covalent association of C3b with C4b within C5 convenase of the classicalcomplement pathway. J. Exp. Med.. 165: 1494-1507. 1987.Masaki. T., Matsumoto, M.. Yasuda, R., Levine, R. P.. Kitamura, H., and Seya, T. Acovalent dimer of complement C4b serves as a subunit of a novel C5 convelíasethatinvolves no C3 derivatives. J. Immunol.. 147: 927-932, 1991.

Sims. P. J.. and Wiedmer, T. The response of human platelets to activated componentsof the complement system. Immunol. Today, 12: 338-342, 1991.Schönermark, S., Filsinger. S., Berger, B., and Hänsch,G. M. The C8-binding proteinof human erythrocytes: interaction with the components of the complement-attackphase. Immunology, 63: 585-590, 1988.

36. Lachmann. P. J. The control of homologous lysis. Immunol. Today. 12: 312-315.

1991.37. Campbell, A. K.. and Morgan B. P. Monoclonal antibodies demonstrate protection of

polymorphonuclear leukocytes against complement attack. Nature (Lond.), 317: 164-

166, 1985.38. Ramm, L., Whitlow. M.. Koski. C., Shin, M., and Mayer, M. Elimination of com

plement channels from the plasma membranes of U937, a nucleated mammalian cellline: temperature dépendanceof the elimination rate. J. Immunol.. ill: 1411-1415.

1983.39. Morgan. B. P.. Danken, J. R.. and Esser. A. F. Recovery of human neutrophils from

complement attack: removal of the membrane attack complex by endocytosis andexocytosis. J. Immunol., 138: 246-253. 1987.

40. Maison. C. M., Villiers, C. L., and Colomb. M. G. Secretion, cleavage and binding ofcomplement component C3 by the human monocytic cell line U937. Biochem. J..261: 407^413, 1989.

41. Ramos. O. F., Algarra, I., Sarmay, G., Yefenof, E., Gergely, J., and Klein, E. Lymphocytes stimulated by allogeneic B cell lines cleave the third component of complement and fix C3 fragments. Their nonspecific lytic capacity is elevated againstcomplement receptor type 2-carrying targets. J. Immunol., 142: 217-223, 1989.

599

Research. on August 20, 2019. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from

1993;53:592-599. Cancer Res   Markus W. Ollert, Joseph V. Kadlec, Eugene C. Petrella, et al.  

p65Protease Melanoma Cells Expressing the C3-cleaving Membrane Molecular Basis of Complement Resistance of Human

  Updated version

  http://cancerres.aacrjournals.org/content/53/3/592

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/53/3/592To request permission to re-use all or part of this article, use this link

Research. on August 20, 2019. © 1993 American Association for Cancercancerres.aacrjournals.org Downloaded from