Ascitic complement system in ovarian cancer

Ovarian cancer spreads intraperitoneally and forms fluid, whereby the diagnosis and therapy often become delayed. As the complement (C) system may provide a cytotoxic effector arm for both immunological surveillance and mAb-therapy, we have characterised the C system in the intraperitoneal ascitic fluid (AF) from ovarian cancer patients. Most of the AF samples showed alternative and classical pathway haemolytic activity. The levels of C3 and C4 were similar to or in the lower normal range when compared to values in normal sera, respectively. However, elevated levels of C3a and soluble C5b-9 suggested C activation in vivo. Malignant cells isolated from the AF samples had surface deposits of C1q and C3 activation products, but not of C5b-9 (the membrane attack complex; MAC). Activation could have become initiated by anti-tumour cell antibodies that were detected in the AFs and/or by changes on tumour cell surfaces. The lack of MAC was probably due to the expression of C membrane regulators CD46, CD55 and CD59 on the tumour cells. Soluble forms of C1 inhibitor, CD59 and CD46, and the alternative pathway inhibitors factor H and FHL-1 were present in the AF at concentrations higher than in serum samples. Despite the presence of soluble C inhibitors it was possible to use AF as a C source in antibody-initiated killing of ovarian carcinoma cells. These results demonstrate that although the ovarian ascitic C system fails as an effective immunological surveillance mechanism, it could be utilised as an effector mechanism in therapy with intraperitoneally administrated mAbs, especially if the intrinsic C regulators are neutralised.

Membrane-bound complement regulatory proteins inhibit complement activation by an immunotherapeutic mAb in a syngeneic rat colorectal cancer model

MAb-mediated immunotherapy offers a potential tool for destroying metastasizing colorectal tumor cells. Promising results have been obtained by using xenograft models. However, overexpression of membrane-bound complement regulatory proteins (mCRP) impedes complement-mediated destruction of tumor cells in vitro. mCRP operate in a species selective manner. Therefore a syngeneic animal model is needed to investigate the contribution of mCRP in mAb-mediated immunotherapy. Here we present a syngeneic rat colorectal carcinoma model, which fulfills the conditions necessary to investigate the effect of mCRP expression on mAb-mediated immunotherapy of metastases of solid tumors.CC531 rat colorectal cancer cells were injected subcapsularly into the liver of syngeneic WAG/Rij rats. Four mAb (MG1(IgG2a), MG2(IgG2a), MG3(IgG3) and MG4(2a)(IgG2a)) directed against CC531 cells, were tested for their complement activating abilities in vitro and tumor homing capacities in vivo. Only MG4(2a) was found to activate complement in vitro and home to the tumor cells in vivo. This mAb induced C3-deposition and C-mediated lysis of CC531 cells in vitro when the effects of the C-inhibitors Crry/p65 and CD59 were neutralized. This implies an important role for these mCRP in restricting the effector functions of tumor-associated mAb on these cells. Although C activation could be induced by MG4(2a) in situ on tumor tissue sections, no deposition of C3 could be found on the tumor cells positive for MG4(2a) in vivo. This suggests that complement activation in vivo was inhibited by mCRP. The results indicate the suitability of this syngeneic animal model for studying the effects of mAb immunotherapy. However, the effect of mCRP on tumor cells need to be overcome, e.g. by the use of mAb against tumor antigens and mCRP.

Secretion of soluble complement inhibitors factor H and factor H-like protein (FHL-1) by ovarian tumour cells

We observed that the soluble complement regulators factor H and factor H-like protein were abundantly present in ascites samples as well as in primary tumours of patients with ovarian cancer. RT-PCR and immunoblotting analyses showed that the two complement inhibitors were constitutively produced by the ovarian tumour cell lines SK-OV-3 and Caov-3, but not PA-1 or SW626 cells. The amounts of factor H-like protein secreted were equal to those of factor H. This is exceptional, because e.g. in normal human serum the concentration of factor H-like protein is below 1/10th of that of factor H. In ascites samples the mean level of factor H-like protein (130+/-55 microg ml(-1)) was 5.5-fold higher than in normal human serum (24+/-3 microg ml(-1)). Ovarian tumour cells thus preferentially synthesise factor H-like protein, the alternatively spliced short variant of factor H. The tumour cells were found to bind both (125)I-labelled factor H and recombinant factor H-like protein to their surfaces. Surprisingly, the culture supernatants of all of the ovarian tumour cell lines studied, including those of PA-1 and SW626 that did not produce factor H/factor H-like protein, promoted factor I-mediated cleavage of C3b to inactive iC3b. Subsequently, the PA-1 and SW626 cell lines were found to secrete a soluble form of the membrane cofactor protein (CD46). Thus, our studies reveal two novel complement resistance mechanisms of ovarian tumour cells: (i) production of factor H-like protein and factor H and (ii) secretion of soluble membrane cofactor protein. Secretion of soluble complement inhibitors could protect ovarian tumour cells against humoral immune attack and pose an obstacle for therapy with monoclonal antibodies.

Complement-mediated killing of microtumors in vitro

Complement-mediated lysis of cancer cells growing in three-dimensional aggregates involves factors that are not associated with the killing of cells in suspension. We have used multicellular tumor spheroids established from breast carcinoma (T47D) and ovarian teratocarcinoma (PA-1) cell lines as models to study complement-mediated destruction of micrometastases and small solid tumors. We found that significant killing of microtumors treated with an antitumor antibody and a specific monoclonal antibody (YTH53.1) against the complement lysis inhibitor protectin (CD59) started to occur after a 1 to 2-hour lag phase. After an overnight incubation, the microtumors became totally infiltrated by the YTH53.1 monoclonal antibody and C1q, whereas C3 and C5b-9 penetrated as a frontier to the peripheral cell layers. A 51Cr release assay showed that during a 24-hour pulsed treatment with complement, 33% of cells in the spheroids were killed, and the average tumor volume decreased by 28%. According to propidium iodide staining, complement exposure resulted in killing and peeling off of the outermost tumor cells.

Expression and function of the complement membrane attack complex inhibitor protectin (CD59) in human prostate cancer

Protectin (CD59) inhibits homologous complement-mediated cytolysis by preventing formation of the membrane attack complex at the point of insertion and polymerization of C9 into cell membranes. The present study investigated the expression and function of CD59 on human prostatic tumor cells in situ and on 5 human prostate cell lines in vitro originating from either metastatic tumors or benign prostate hypertrophy epithelial cells. Immunohistochemical staining of prostate carcinoma tissue with monoclonal antibody (MAb) MEM43 revealed weak to moderately strong expression of CD59 by prostate glandular epithelial cells. Flow cytometry with MEM43 demonstrated that the 5 prostate cell lines expressed different relative quantities of CD59. Indirect immunofluorescence analysis revealed uniform membrane staining of DU145 and PC3 cell lines with no membranous granularity in the staining pattern. Western immunoblots with MAb BRIC 229 showed that PC3 and DU145 cells express CD59 with a m.w. of 18-25 kDa. Treatment of DU 145 and PC3 cells with phosphatidylinositol-specific phospholipase C caused a significant decrease of CD59 expression indicating that the CD59 expressed by prostate cancer cells is anchored to the cell membrane via a glycosylphosphatidylinositol (GPI) linkage. PC3 and DU145 cells were completely resistant to human complement-mediated cytolysis but became sensitive to killing in the presence of the CD59-neutralizing MAb YTH53.1. We conclude that malignant and benign human prostate cells express CD59 that is GPI-linked to the cell surface and that CD59 may regulate the immunological response to cancerous prostate cells by protecting the cells from the cytolytic activity of complement.

Complement-regulatory proteins in ovarian malignancies

Ovarian cancer has features that makes it well-suited for MAb adjuvant immunotherapy. Several of the MAbs used in clinical trials mediate cancer cell destruction by activation of complement (C). In this study, therefore, we examined the ability of ovarian-tumor cells to resist C attack. We found that the C regulators membrane cofactor protein (MCP, CD46) and protectin (CD59) were strongly expressed in the tumor cells in all 28 benign and malignant tumors examined. Decay-accelerating factor (DAF; CD55) was more heterogeneously expressed, and only 75% of the tumors exhibited a moderate amount of DAF in the tumor cells. In adenoma cells, CD59 and DAF were preferentially located apically, while in adenocarcinoma cells they were expressed also at the basolateral cell surface. The ovarian-carcinoma cell lines SK-OV-3, Caov-3, SW626 and PA-1 expressed both the 58- and the 68-kDa isoforms of MCP. DAF was present as a glycosyl-phosphatidylinositol(GPI)-anchored 70-kDa glycoprotein. The surface-expression level of DAF varied, and correlated with the vulnerability of the cells to C-mediated lysis. CD59 was expressed as a GPI-linked 19- to 25-kDa protein exhibiting multiple glycosylation variants. The surface expression of CD59 correlated with the amount of the main 1.9 + 2.1-kb CD59 mRNA transcripts. Neutralization of CD59 with an anti-CD59 MAb significantly enhanced C-mediated killing of the cell lines. Low expression of C regulators on the PA-1 teratocarcinoma cell line was associated with high sensitivity to C lysis. Thus, the expression of C regulators on malignant ovarian cells may constitute a tumor escape mechanism, and is a critical parameter to be examined when MAb therapy is being considered.

Resistance of ovarian teratocarcinoma cell spheroids to complement-mediated lysis

We have shown previously that it is possible to target complement-mediated killing against cultured ovarian tumour cells in vitro. As malignant ovarian cells usually grow in solid nodules in vivo, we have in the present study examined the effectiveness of complement killing against ovarian teratocarcinoma cells (PA-1) growing in three-dimensional tumour microspheroids (TMSs). Our study shows that PA-1 cells growing in TMSs are less susceptible to complement-mediated killing than cells growing in monolayer cultures, even after neutralization of protectin (CD59), the main inhibitor of complement lysis. Cells in suspension and cells growing in TMSs showed a similar expression of membrane co-factor protein (MCP, CD46) and CD59. Decay-accelerating factor (DAF, CD55) was not detected on the surface of cells in suspension, but appeared focally on the outermost cell layers of the TMSs. Complement-activating antibodies bound to all PA-1 cells in suspension but only to the most peripherally located cells in TMSs, even though the target antigens were similarly expressed in the two systems. Antibody-induced complement activation on PA-1 cells in suspension led to C3 and C5b-9 deposition on most cells, while C3 and C5b-9 were only found on the outermost layers of the TMSs. The increased complement resistance of tumour cells growing in three-dimensional spheroids is partly because of an insufficient penetration of antibodies and complement into the TMSs. TMSs are a useful model for the development of more efficient ways to kill malignant cells in micrometastases with monoclonal antibodies and complement.

Analysis of the recognition mechanism of the alternative pathway of complement by monoclonal anti-factor H antibodies: evidence for multiple interactions between H and surface bound C3b

The ability of the alternative pathway of complement to discriminate targets as either activators or non-activators is mediated by different binding properties of factor H to surface-associated C3b molecules. In the present study we have probed the interaction between H and C3b using five anti-H mAb. The binding sites of the mAb were mapped by Western blotting using both recombinant and trypsin-generated H fragments. Two mAb bound to CCP1 (90X, 196X), two to CCP5 (MRC OX24, 86X) and one to CCP8-15a (131X). At a molar ratio 2:1 of 125I-H:mAb all tested mAb enhanced binding of H to both activator- and non-activator-bound C3b. At higher concentrations two mAb had an inhibitory effect on H binding to surface-associated C3b (OX24, 131X). Thus the mAb 131X inhibits H binding to surface-bound C3b but unlike OX24 it does not bind to the previously described C3b binding site within or near CCP4-5. These results indicate that there is an additional interaction site on factor H for surface-bound C3b.

Expression of complement membrane regulators membrane cofactor protein (CD46), decay accelerating factor (CD55), and protectin (CD59) in human malignant gliomas

Gliomas are malignant brain tumors, which, despite recent progress in surgical and radiological treatment, still have a poor prognosis. Since gliomas apparently resist immunological clearance mechanisms, we became interested in examining bow gliomas resist killing by the human complement system. The resistance of human cells to complement-mediated damage is, in large part, mediated by specific inhibitors of complement:membrane cofactor protein (CD46), decay-accelerating factor (CD55), and protectin (CD59). In the present study we examined the expression of complement regulators in 14 human glioma tumors and in 7 glioma cell lines (U251, U87, HS683, U373, U138, U118, and H2). Protectin was found to be strongly expressed by all glioma tumors and cell lines. Northern blotting analysis demonstrated the typical pattern of four to five protectin mRNAs in the glioma cells. Except for blood vessels, the expression of decay-accelerating factor was weak or absent in the tumors in situ, whereas in the cell lines its expression varied, ranging from negative to intermediate. Membrane cofactor protein was moderately expressed by all the cell lines but only weakly in the tumors. Cell-killing experiments demonstrated that the glioma cell lines were exceptionally resistant to C-mediated lysis. Five of the seven cell lines (U373, HS683, U118, U138, and H2) resisted complement lysis under conditions where most other cell lines were sensitive to killing. Neutralization experiments using specific monoclonal antibodies indicated that protectin was functionally the most important complement regulator in the glioma cells. The killing of the U87 and U251 cells could be significantly increased by a blocking anti-protectin monoclonal antibody, whereas for the other cell lines only moderate or no response was observed. The H2 cell line resisted killing by all antibodies and by complement. These results show that protectin is the most important complement regulator on human glioma cells. The exceptional complement resistance of some glioma cell lines suggests that they may utilize other, hitherto less well characterized, mechanisms to resist complement killing.

Shedding and enrichment of the glycolipid-anchored complement lysis inhibitor protectin (CD59) into milk fat globules

Protectin (CD59) is a glycolipid-anchored inhibitor of the membrane attack complex (MAC) of human complement (C) that protects blood cells, endothelial cells and various epithelial cells from C-mediated lysis. Because of its activities protectin is a candidate molecule for use in the treatment of paroxysmal nocturnal haemoglobinuria or conditions where MAC causes tissue damage. Soluble, phospholipid-free forms of protectin have been isolated from human urine and produced in recombinant form, but they have only a relatively weak C lysis-inhibiting activity. In the present study we have looked for functionally active protectin in human breast milk. Milk is rich in fat droplets, milk fat globules (MFG), that are enveloped in a plasma membrane derived from secretory cells of the mammary gland. The membranes of MFG contain a variety of glycoproteins expressed by the mammary epithelial cells. Both immunofluorescence and immunoblotting analysis demonstrated that protectin was strongly expressed on human MFG. In sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis, MFG protectin (CD59M) appeared as distinct bands with apparent molecular weights of 19,000-23,000 MW, similar to protectin extracted from MCF7 breast carcinoma cells. CD59M in breast milk was functionally active and had a glycophospholipid anchor, as judged by its ability to incorporate into guinea-pig erythrocytes and inhibit their lysis by human complement. These results indicate that functionally active protectin becomes enriched in MFG and imply that secretion of glycophospholipid-anchored molecules, e.g. into cow milk and colostrum, could be exploited as a means of producing bioactive molecules that need to be targeted into cell membranes.

Expression and function of the complement membrane attack complex inhibitor protectin (CD59) on human breast cancer cells

BACKGROUND

Normal human cells resist the lytic activity of homologous complement (C) by expressing inhibitory molecules on their cell membranes. Recently, it has become increasingly evident that information on C inhibitors on malignant tumor cells is crucial before considering any immunotherapeutic attempts with C-activating antibodies. As one of the most potent inhibitors of C lysis is protectin (CD59), we have examined its expression and function on human breast cancer cells.

EXPERIMENTAL DESIGN

Immunofluorescence microscopy was used to detect protectin expression on solid breast tumor samples (N = 12). Using immunoaffinity chromatography, protectin was isolated from the membranes of cultured MCF7 and T47D breast cancer cells. The purified proteins were incorporated into heterologous cells to study their C inhibitory activities. The reactivity of tumor cell protectins with terminal C complexes was examined by sucrose density ultracentrifugation analysis. A chromium release assay was used to study the effects of protectin neutralization on the sensitivity of MCF7 and T47D cells to C-mediated cytotoxicity.

RESULTS

Protectin was found to be strongly expressed by all human breast cancer tumors examined. The affinity-purified protectins had a glycophosphoinositollipid anchor and migrated in sodium dodecyl sulfate-polyacrylamide gel electrophoresis as glycosylated smears of 19 to 25 kilodaltons. Protectin isolated from T47D cells bound to nascent C5b-9 complexes generated in human sera and inhibited C lysis of guinea pig erythrocytes when incorporated into their cell membranes. C-mediated killing of breast cancer cells could be significantly enhanced after treatment of the cells with F(ab')2 fragments of the anti-protectin monoclonal antibody YTH53.1.

CONCLUSIONS

Human breast cancer cells resist C membrane attack by expressing protectin on their cell membranes. Neutralization of protectin on the surface of the tumor cells increases their sensitivity to C lysis.

Targeted neutralization of the complement membrane attack complex inhibitor CD59 on the surface of human melanoma cells

Major problems in the immunotherapy of human tumors with complement-activating monoclonal antibodies (mAb) are (i) inherent resistance of tumor cells to complement cytolysis and (ii) a possible undiscriminatory attack against normal cells. In the present study we have developed a procedure to simultaneously direct the complement membrane attack complex and neutralize its inhibitor CD59 (protectin) on human melanoma cells in vitro. G361 melanoma cells were selectively recognized in heterogenous cell mixtures by a complement-fixing mAb (R24) against the tumor cell GD3-ganglioside. Biotinylated anti-CD59 mAb (YTH53.1) was directed to the tumor cells with a high-affinity biotin-avidin bridge using a proportion of R24 as a biotinylated targeting mAb and avidin as a linker. Biotinylated anti-CD59 mAb lost its ability to activate complement, but retained its CD59-neutralizing activity. Thus, it was possible to avoid nonspecific lysis of surrounding erythrocytes and endothelial cells and direct the CD59-neutralizing effect to the tumor cells. As a result the tumor cells were efficiently killed by R24 plus complement while the bystander cells remained viable. These results suggest that it is possible to target an unrestricted complement membrane attack against GD3- and CD59-positive melanoma cells.