Clinical hypothermia temperatures increase complement activation and cell destruction via the classical pathway

Pooled normal human serum (NHS) was derived from the blood of healthy human volunteers obtained with written informed consent in accordance with an Institutional Review Board approved protocol (IRB 02-06-EX-0216, Eastern Virginia Medical School).

Sheep erythrocytes were purchased from MP Biomedicals, LLC. Anti-sheep-erythrocyte-IgM was purchased from Rockland Incorporated. Sheep erythrocytes were incubated with antibody at 30°C to generate antibody-sensitized cells (EA), as previously described [23]. All antibody-sensitization was performed at 30°C. Standard complement buffers were prepared for hemolytic assays: GVBS⁺⁺ (Veronal buffered saline, 0.1% gelatin, 0.15 mM CaCl₂, and 1.0 mM MgCl₂), Mg-EGTA-GVBS (Veronal buffered saline with 5 mM MgCl₂ and 8 mM EGTA), and GVBS^{- -} (Veronal buffered saline, 0.1% gelatin, 0.01 M EDTA), Veronal Buffered Saline (10 mM Barbital, 145 mM NaCl, pH 7.4). Peptide Inhibitor of Complement C1 (PIC1) version AcPA, described elsewhere [24], was synthesized by New England Peptide (Gardner, MA). Lyophilized PIC1 was reconstituted in DMSO. Pooled normal human serum (NHS) was prepared as previously described [25]. C8-depleted human serum and purified C1q were purchased from CompTech (Tyler, TX).

Antibody-initiated complement hemolytic assays were performed by incubating 100 μL EA with 0.2% NHS in GVBS⁺⁺. Hemolytic assays were performed for one hour at temperatures ranging from 0°C to 37°C. A separate assay was performed with more clinically relevant temperatures of 31°, 33°, 35°, 37°, 39°, and 41°C. EA were also incubated with water or buffer as controls. Hemolysis reactions were quenched with 4 mL GVBS^{- -} and the supernatants recovered after sedimentation. Hemolysis was quantified by spectrophotometry of the supernatants at 412 nm and normalized as previously described [23]. Alternative complement pathway hemolytic assays were performed by incubating rabbit erythrocytes (CompTech, Tyler, TX) with 4% NHS in Mg-EGTA-GVBS at temperatures ranging from 0°C to 37°C. A separate assay was performed with more clinically relevant temperatures of 31°, 33°, 35°, 37°, 39°, and 41°C. Reactions were quenched with 2 mL GVBS^{- -}, sedimented and measured as above.

Samples were generated by incubating 1% NHS with 400 μg/ml heat-aggregated IgG in GVBS⁺⁺ at each temperature for one hour. Heat-aggregated IgG was generated from purified human IgG (Gammagard, Baxter) as described elsewhere [23]. The reaction was stopped with EDTA-GVBS^{- -} and the samples were measured by C5a ELISA (R&D Systems). Background values from controls of NHS and heat-aggregated human IgG incubated on ice for one hour were subtracted.

C3/C4 samples were generated by incubating C8-deficient serum (to prevent hemolysis), 1% final, with 0.25 mL EA in a total of 0.75 mL of GVBS⁺⁺ for one hour at each temperature. After incubation, the samples were washed twice with water to remove hemoglobin, and the resulting membrane bound C3/C4 was stripped using 25 mM methylamine. These samples were quantitated by total C3 ELISA, iC3b ELISA, and C4 ELISA as previously described [26]. C3 and C4 Western blot analysis was performed as described elsewhere [26].

A Western blot to assess C1s activation by heat aggregated IgG has been previously described [27, 28] One microliter of partially purified C1 (0.2 mg/ml, Complement Technologies, Inc.), 5 μl of heat-aggregated IgG (1:250 dilution from 50 μg/ml stock), and 5 μl of PBS were combined on ice and samples incubated at the following temperatures: 31, 33, 35, 37, 39 and 41°C for 0, 30, 45, 60 and 90 minutes. After incubation, 4 μl of loading buffer was added and each sample was boiled. Gel electrophoresis on a 10% SDS-PAGE gel was performed at 140 V for 60 minutes. The gel was then transferred to a nitrocellulose membrane at 200 mA for 60 minutes. The membrane was washed twice for five minutes in PBS buffer and then blocked overnight in 5% NFDM-PBS. The next day, the membrane was probed with primary antibody, Goat anti-C1s (Quidel), at a dilution of 1:2000 in 5% NFDM-PBS/0.1% Tween-20 for one hour at room temperature. The membrane was washed 3X for 10 minutes with PBS/0.1% Tween-20. Then, the membrane was probed with secondary antibody, Donkey anti-goat IR680 (Li-cor Biosciences), at a dilution of 1:10,000 in 5% NFDM-PBS/0.1% Tween-20 for one hour at room temperature. After washing the membrane 3X with PBS/0.1% Tween-20, it was imaged on an Odyssey imager to determine the amounts of the C1s heavy and light chains characteristic of activated C1s relative to the proenzyme species.

C1 either from NHS (0.33% final) or C8-deficient serum (CompTech), or purified C1q (Comp Tech) at 0.24 μg/ml final, was combined with 100 μl of sensitized sheep erythrocytes in a total volume of 0.75 ml of GVBS++ for one hour at the various temperatures. These cells were then washed twice with water and the resulting cell pellet was dissolved with 0.5% NP40. Membrane bound C1q was analyzed by a quantitative Dot Blot assay. Pure C1q was titrated onto PVDF membrane, along with the samples, via a Dot Blot apparatus. The membrane was blocked with 3% BSA, probed with a goat anti-C1q antibody (Comp Tech), and then an HRP-labeled rabbit anti-goat secondary antibody (Sigma-Aldrich). Bound antibody was detected by enhanced chemiluminescence. Using the Quantity One software (Bio-Rad), grey scale values were assigned to the pure C1q titration and the samples so that a linear regression could be used to quantify the amount of C1q in each sample.

Peptide Inhibitors of complement C1 (PIC1) specifically inhibit C1 activation, as previously described [24, 28]. Here we use the AcPA form (IALILEPICCQERAA) [24] of PIC1. Preparations of 16.5% normal human serum (NHS) in GVBS⁺⁺ were incubated with 0.98 mM, 0.82 mM, 0.65 mM, 0.48 mM or 0.32 mM PIC1, DMSO vehicle control, or buffer control at 4°C for 1 hour. Samples were further diluted 1:10 in GVBS⁺⁺. Hemolysis reactions were then performed with 100 μL of NHS sample plus 550 μL of additional GVBS⁺⁺ and 100 μL of EA over 1 hour at either in 31° or 37°C. Controls and samples were processed and measured by spectrophotometry at 412 nm.

In order to evaluate antibody-initiated complement-activation at therapeutic hypothermia temperatures, we tested antibody-sensitized sheep erythrocytes (EA) incubated in normal human serum at 31°C-41°C and measured hemolysis. In this CH50-type assay, hemolysis of antibody-sensitized erythrocytes in human serum differed significantly as a function of temperature (ANOVA P < 0.01). There was 9-fold more hemolysis at 31°C compared with 41°C (t-test P < 0.01) and 2 fold more hemolysis at 31°C compared with 37°C (t-test P = 0.01) (Figure 1A). These results show that antibody-initiated complement activation increased at therapeutic hypothermia temperatures. Given that complement activation is severely inhibited at near 0°C temperatures, we then evaluated complement activation for temperatures ranging from 0°C-37°C. Antibody-initiated complement activation in human serum increased as temperature decreased from 37°C to 24°C until dramatically decreasing at 13°C (Figure 1B). In these assays a similar 2-fold increase in antibody-initiated complement activation was noted comparing 31°C with 37°C (t-test P < 0.01). From 0°C – 13°C complement activation was minimal.We then tested whether similar effects would be observed for the alternative complement pathway over therapeutic hypothermia temperatures. In an AP50-type alternative complement pathway assay, temperature did not alter complement-mediated cell lysis at therapeutic hypothermia temperatures (ANOVA P = 0.45) (Figure 1C). Testing alternative pathway activation from 0°C – 37°C demonstrated that activation was significantly inhibited at 24°C and below (Figure 1D). These results show that antibody-initiated complement activation increased at clinical hypothermia temperatures compared with euthermia, suggesting that antibody-initiated complement activation occurring during reperfusion at hypothermic temperatures could potentially increase complement-mediated tissue damage.

Complement-mediated hemolysis results from pores generated by the membrane attack complex (MAC), suggesting that increased terminal complement cascade activation was occurring at lower temperatures. In order to further evaluate activation of the terminal complement cascade we measured C5a generation in response to a different antibody-initiated complement activator, heat-aggregated IgG. C5a is also an extremely potent anaphylatoxin important for phagocytic cell recruitment and activation [29] and implicated in contributing to tissue damage in many inflammatory diseases processes [30]. Heat-aggregated IgG incubated with normal human serum demonstrated that C5a generation differed significantly as a function of temperature (ANOVA P < 0.01). There was a >2-fold increase in C5a generation at 31°C compared with 37°C (t-test P =0.05) (Figure 2A). These data show that at clinical hypothermia temperatures terminal complement cascade activation increased, consistent with the CH50-type assay, and suggests that increased C5a generation at these temperatures could increase neutrophil recruitment and activation enhancing inflammation and host tissue damage during reperfusion.

C3 is the central component of the complement cascades and its activation results in generation of the anaphylatoxin C3a and opsonization of targets with the C3-fragments, C3b and iC3b. We tested C3 activation and opsonization with C3b/iC3b by incubating EA with C8-deficient serum (to prevent hemolysis). Though total C3-fragment opsonization as a function of temperature was not statistically significant (ANOVA P = 0.5) (Figure 2B), cell membrane bound iC3b differed significantly as a function of temperature (ANOVA P < 0.01), with a 2-fold increase at 31°C versus 37°C (t-test P = 0.04) (Figure 2C). Western blot analysis of membrane-bound C3-fragments confirmed that iC3b was the predominant form and suggested increased iC3b deposition at lower temperatures (Figure 2D) consistent with ELISA data. These data show that C3 activation and iC3b opsonization increased at clinical hypothermia temperatures, suggesting that these temperatures could increase opsonization of host cells (e.g. endothelial cells) during reperfusion targeting them for phagocyte attack.

In order to specifically evaluate the classical complement pathway, we tested C4 activation by incubating EA with C8-deficient serum. C4-fragment binding differed significantly as a function of temperature (ANOVA P = 0.04). There was an almost 2-fold increase in C4 generation at 31°C compared with 37°C (t-test P = 0.05) (Figure 2E). Western blot analysis demonstrated a mixture of C4b, iC4b and C4d fragments (Figure 2F) and suggested increased C4-fragment binding at 31°C. These findings show that classical pathway activation is increased at therapeutic hypothermia temperatures and suggests that C1 activation, which cleaves C4, is also increased at these temperatures.

C1 is a complex molecule consisting of a pattern (e.g., IgG/M) recognition component C1q, which associates with the serine protease tetramer C1r-C1s-C1s-C1r. When the C1q portion of the molecule binds to IgG/M, it causes auto-activation of the enzymatic tetramer leading to activation of C4 and the classical pathway cascade. Therefore, increased activation of C1 at lower temperatures could be caused by increased C1q binding, or increased enzymatic activity, or both. C1 enzymatic activation was tested at increasing temperatures over a 90 minute period using heat-aggregated IgG incubated with partially purified C1. C1s activation was assayed by the conversion of precursor C1s to its activated heavy and light chain forms, as previously described [27, 28]. C1s activation decreased at therapeutic hypothermia temperatures, suggesting decreased enzymatic activity at lower temperatures (Figure 3).C1q and C1 binding was tested by incubating EA with purified C1q, or normal human serum, or C8-deficient serum. In each case, C1/C1q binding to EA showed a consistent trend of increasing at lower temperatures (Figures 4A, B, and C), suggesting that increased classical complement pathway activation is likely mediated by increased C1 binding at therapeutic hypothermia temperatures. These findings may also explain the lack of hypothermia-enhanced alternative pathway activation where initiation is not mediated by a pattern recognition molecule (e.g. C1q) binding event.

In order to further elucidate the role of C1 in antibody-initiated hypothermia-enhanced complement activation, we tested a specific inhibitor of C1 activation, PIC1, (Peptide Inhibitor of complement C1) [23, 24, 27, 28]. In the CH50-type assay, PIC1 successfully inhibited hypothermia-enhanced complement activation at 31°C in a dose-dependent manner by up to 60% compared with untreated control (ANOVA P = < 0.01) (Figure 5). At 0.98 mM PIC1 inhibited hypothermia-enhanced (31°C) complement-mediated cell lysis to a level not statistically different from that for euthermia (37°C) (t-test P = 0.44). These results suggest that hypothermia-enhanced antibody-initiated complement activation can be blocked using a specific inhibitor of C1 activation.