PROTECTION OF PORCINE ENDOTHELIAL CELLS FROM COMPLEMENT-MEDIATED CYTOTOXICITY BY THE HUMAN COMPLEMENT REGULATORS CD59, C1 INHIBITOR, AND SOLUBLE COMPLEMENT RECEPTOR TYPE 1

Analysis in a Pig-to-Human In Vitro Model Relevant to Hyperacute Xenograft Rejection

Abstract

Inhibition of complement activation is considered a prerequisite to overcome hyperacute xenograft rejection. In the present study, we investigated the efficacy of C1 inhibitor (C1 inh) and recombinant soluble complement receptor type 1 (rsCR1) to protect xenogeneic cells against complement-mediated cytotoxicity in an in vitro xenotransplantation model. The addition of the soluble complement regulators to human serum led to a dose-dependent inhibition of complement-mediated destruction of aortic porcine endothelial cells (PEC). On a molar base, rsCR1 was more efficient than C1 inh. Transfection of PEC with cDNA of human CD59 resulted in several clones where protection against complement-mediated cell destruction correlated with the expression level of the inhibitor. Addition of low concentrations of C1 inh and rsCR1 to a CD59 (human)-positive PEC clone, expressing a suboptimal level of the membrane-bound regulator, resulted in a significant improvement of protection against complement-mediated cell destruction.

Complement-mediated hyperacute graft rejection leading to organ failure within minutes to a few hours is probably the most critical barrier to successful xenotransplantation ⁽¹⁾. Complement depletion by cobra venom factor with subsequent prolongation of graft survival provided the first evidence that complement inhibition may be an effective means to control hyperacute xenograft rejection ⁽²⁾. At present, support of the physiological regulation of the complement system appears to be most promising. In various animal models of xenotransplantation, pretreatment of the recipient with soluble complement receptor type 1 significantly improved graft survival ^(3-5). Another candidate molecule may be C1 inhibitor (C1 inh^*), a major plasma inhibitor of the complement and contact systems. C1 inh, which has long been used for the treatment of the hereditary angioedema, has recently been introduced as substitution therapy in patients with vascular leak syndrome, severe sepsis, and myocardial ischemia/reperfusion injury^{(6, 7)}. The transfer of human membrane-bound complement-regulatory molecules, such as decay-accelerating factor (CD55), membrane cofactor protein (CD46), and CD59, to the donor organ has been suggested as an alternative strategy to protect the xenograft from the attack of the recipient's complement system ^(8-10). Because pigs are considered to be suitable organ donors for humans, we investigated the efficacy of human recombinant soluble complement receptor type 1 (rsCR1; ¹¹) and C1 inh, either alone or in combination with the membrane-associated complement control protein CD59, to protect aortic porcine endothelial cells (PEC) against complement-mediated cytotoxicity of human serum. Endothelial cells were chosen for our experiments because they represent the primary target cells involved in hyperacute graft rejection.

Endothelial cells were isolated from porcine aorta and cultured as described recently ⁽⁹⁾. For the experiments, cells between passages 3 and 15 were used. Human rsCR1 was kindly provided by Dr. James Levin, T-Cell Sciences Inc., Needham, MA. C1 inh (Behrinert) was purchased from Behring, Marburg, Germany. Incremental amounts of C1 inh and rsCR1 were first added to normal human serum (NHS) as a source of xenogeneic natural antibodies and complement, before incubation with PEC. After 2 hr, cell lysis was quantified by measuring lactate dehydrogenase in the supernatant, as described previously ⁽¹²⁾. In the absence of complement inhibitors, 20% NHS caused 40-60% specific lysis of the cells. Incubation with heat-inactivated serum with or without inhibitors only slightly increased the level of spontaneously released lactate dehydrogenase. Addition of each of the two inhibitors caused a dose-dependent reduction in cytotoxicity (Figs. 1 and 3). Twenty-eight micromoles of C1 inh per liter of undiluted serum resulted in complete protection from complement-mediated lysis (Fig. 1). In the presence of 1-3 μM C1 inh, which corresponds to the clinically administered dose ^{(6, 7)}, a 40-60% inhibition was achieved. In comparable experiments performed by Dalmasso and Platt⁽¹³⁾, 40 μM C1 inh were required to obtain 50% inhibition of lysis. This may be explained by a different functional activity of the C1 inh preparation used.

Complete protection of PEC by rsCR1 (Fig. 3) was observed at a concentration of 1.5 μM, corresponding to 300 μg/ml, a dose within the concentration range used in animal xenotransplantation models^(3-5). It has been shown that porcine hearts perfused ex vivo with human blood, containing 300 μg rsCR1/ml, maintained their function for up to 4 hr, compared with 34 min in the absence of the inhibitor ⁽⁵⁾. A single intravenous bolus of rsCR1 (15 mg/kg) administered to cynomolgus monkeys immediately before reperfusion prolonged survival of a porcine cardiac xenograft from 1 hr to 48-90 hr⁽⁵⁾. In a recent animal study, we were able to extend survival time of guinea pig kidney grafts in rats from about 5 min up to 20 hr upon pretreatment of the recipients with 50 mg/kg rsCR1⁽³⁾.

In our cell culture experiments, 50% protection from lysis could be achieved with about 0.05 μM rsCR1 (Fig. 3), indicating that-on a molar base-rsCR1 was about 60-fold more efficient than C1 inh. However, despite the higher potency of rsCR1 in the cell culture model, the application of C1 inh in vivo may offer certain advantage: C1 inh is not only the major plasma inhibitor of the classical pathway of the complement system, but also of importance in the regulation of the coagulation system, which is known to contribute to hyperacute graft rejection. C1 inh selectively interferes with the classical pathway of complement, thereby leaving the alternative pathway available to protect the immunosuppressed graft recipient against bacterial infections. Selective inhibition of C1 may be sufficient in species combinations such as pig/human, where complement is mainly activated via the classical complement pathway.

CD59-positive cell clones were produced by transfection of PEC with CD59 cDNA subcloned into the expression vector pHβapr-1-neo, as described recently ⁽¹²⁾. CD59 transfectants were selected in culture medium containing 0.6 mg/ml G418 (Gibco, Eggenstein, Germany) and cloned by limiting dilution. The CD59 level was determined by quantitative cytofluorometric analysis (Quifikit; Dako, Hamburg, Germany) according to the manufacturer's protocol. For this purpose, 5×10⁵ cells were incubated for 30 min on ice with saturating amounts of monoclonal anti-CD59 antibody MEM-43 (mouse IgG2a; Cell Systems, Remagen, Germany). Nonimmune mouse IgG2a (Sigma, München, Germany) served as control. After intensive washing, fluorescein isothiocyanate-conjugated goat anti-mouse IgG (Dako) was applied. In parallel, calibration beads coated with different, well-defined quantities of mouse IgG that mimic cells with different antigen densities were treated with saturating amounts of the second antibody. After 30 min on ice, cells and beads were washed again twice and analyzed in a fluorescence-activated cell sorter (FACScan; Becton Dickinson, Mountain View, CA). The mean fluorescence intensity of each population of calibration beads was plotted versus the number of mouse monoclonal IgG per microbead. From the mean fluorescence intensity of CD59-reactive cells and negative control samples the antigen density was determined by interpolation of the calibration curve. We obtained various PEC clones where CD59 surface expression level ranged from 5×10⁴ to 1×10⁶ molecules/cell.

The susceptibility of these clones to complement-mediated lysis was investigated as described above. As shown in Figure 2, the resistance of these clones to complement-mediated lysis correlates with the amount of human inhibitor on the cell surface. Clones expressing about 1×10⁶ CD59 molecules/cell were completely protected from lysis for all NHS concentrations up to 40%. At a level of 1-2×10⁵ inhibitor molecules CD59-PEC were protected against 20% NHS(Fig. 2), but became susceptible to 40% NHS (data not shown). Less than 1×10⁵ CD59 molecules/cell provided only limited protection from human serum (Fig. 2).

To investigate a possible synergistic effect of soluble and membrane-bound complement inhibitors, we chose experimental conditions where a suboptimal amount of each of the soluble complement regulators resulted in a partial protection of cells. As a target for complement-mediated lysis, a CD59-PEC clone was selected that expressed about 9×10⁴ CD59 molecules on the cell surface (marked by an arrow in Fig. 2). This level provides only limited protection (40%) from complement-mediated lysis under the conditions described above. The addition of 0.04-3 μM C1 inh led to dose-dependent augmentation of the protective effect(Fig. 1). At a concentration of as low as 0.04 μM, which by itself had no effect in experiments with CD59-negative cells(Fig. 1), protection of CD59-PEC was augmented as compared with cells incubated with serum in the absence of C1 inh(Fig. 1). Similar results were obtained with rsCR1(Fig. 3). Applying 2-50 nM rsCR1, complement-mediated lysis of CD59-PEC was diminished in a dose-dependent manner. Almost complete protection was achieved in the presence of 50 nM rsCR1, a concentration that led to only 50% inhibition of cytotoxicity of nontransfected PEC. Even at 2 nM rsCR1, a concentration too low to inhibit complement-mediated lysis of nontransfected PEC, a protective effect exceeding that mediated by human CD59 alone was observed.

Concentrations of the two soluble complement regulators examined were in the range of those applied in vivo and have been shown in clinical trials to be without side effects ^{(6, 14)}. Control of hyperacute xenograft rejection can be improved, if local protection by membrane-associated human complement regulators in donor organs of transgenic animals ^{(8-10, 15)} is augmented by the additional administration of low doses of physiological soluble complement inhibitors. This approach lowers the expression level of membrane-associated complement inhibitors required to be effective.

Even if hyperacute graft rejection may be abrogated by successful complement inhibition, the graft most probably will be exposed to the recipient's cell-mediated immune response. Here the proposed regimen bears the advantage of minimizing the risk of infection associated with an indispensable immunosuppressive therapy.

Figure 1: Protection of PEC against complement-mediated lysis by C1 inh. Nontransfected PEC (□) and CD59-positive PEC (▪) were incubated with 20% NHS. Results refer to neat (100%) serum and expressed as mean ± SEM of ≥3 separate experiments.

Figure 2: Correlation of protection from complement-mediated lysis with the level of human CD59 expressed on CD59-PEC. PEC clones expressing different amounts of CD59 were incubated with 20% NHS. The CD59-PEC clone used in further experiments is indicated by an arrow. Results are expressed as mean of≥3 separate experiments.

Figure 3: Effect of rsCR1 on complement-mediated lysis of nontransfected(□) and CD59-transfected PEC (▪). Results refer to neat (100%) serum and expressed as mean ± SEM of ≥3 separate experiments.