Complement-Mediated Enhancement of Monocyte Adhesion to Endothelial Cells by HLA Antibodies, and Blockade by a Specific Inhibitor of the Classical Complement Cascade, TNT003

Abstract

Background: Antibody-mediated rejection (AMR) of most solid organs is characterized by evidence of complement activation and/or intragraft macrophages (C4d + and CD68+ biopsies). We previously demonstrated that crosslinking of HLA I by antibodies triggered endothelial activation and monocyte adhesion. We hypothesized that activation of the classical complement pathway at the endothelial cell surface by HLA antibodies would enhance monocyte adhesion through soluble split product generation, in parallel with direct endothelial activation downstream of HLA signaling.

Methods: Primary human aortic endothelial cells (HAEC) were stimulated with HLA class I antibodies in the presence of intact human serum complement. C3a and C5a generation, endothelial P-selectin expression, and adhesion of human primary and immortalized monocytes (Mono Mac 6) were measured. Alternatively, HAEC or monocytes were directly stimulated with purified C3a or C5a. Classical complement activation was inhibited by pretreatment of complement with an anti-C1s antibody (TNT003).

Results: Treatment of HAEC with HLA antibody and human complement increased the formation of C3a and C5a. Monocyte recruitment by human HLA antibodies was enhanced in the presence of intact human serum complement or purified C3a or C5a. Specific inhibition of the classical complement pathway using TNT003 or C1q-depleted serum significantly reduced adhesion of monocytes in the presence of human complement.

Conclusions: Despite persistent endothelial viability in the presence of HLA antibodies and complement, upstream complement anaphylatoxin production exacerbates endothelial exocytosis and leukocyte recruitment. Upstream inhibition of classical complement may be therapeutic to dampen mononuclear cell recruitment and endothelial activation characteristic of microvascular inflammation during AMR.

FIGURE 1

Peripheral blood monocyte adhesion to HLA alloserum-stimulated endothelial cells is significantly enhanced in the presence of activated human complement, but not C1q-depleted complement. HAEC were left untreated, treated with negative alloserum (NS) containing no HLA antibodies, or treated with HLA allosera containing antibodies to HLA antigens carried by the endothelial cell for 20 minutes. Primary human monocytes (PBMC-mo) purified from peripheral blood were fluorescently labeled with CFSE, then added to endothelial monolayers in the presence of 25% heat-inactivated human complement (HI C’, black border), intact complement (C', grey fill), or C1q-depleted complement (C1q-dpltd, grey border) and allowed to adhere for 45 minutes. Nonadherent monocytes were removed by washing and monolayers were fixed. Adherent monocytes were visualized by fluorescence microscopy in the FITC channel, and counted in 5 to 8 fields per condition. A, Representative 4x fluorescent fields of CFSE-labeled monocytes (white spots) bound to endothelial monolayers are given. B, Box and whiskers plot shows one representative experiment of monocytes binding to HAEC from donor Y126, treated with negative serum (NS) or 2 HLA allosera (M0546 and J1179). The number of adherent monocytes per field is shown, with error bars demonstrating the range from minimum to maximum, and central line at the median. Dashed line represents the baseline monocyte adherence in the untreated condition with HI C’. C, Scatter dots summarize results from multiple independent experiments (“NS,” n = 4; “alloserum,” n = 12 total unique combinations of monocytes from 3 different donors against endothelial cells from 3 different donors, each stimulated with different allosera). Each data point represents a unique HAEC-serum-monocyte combination. Heat-inactivated complement is shown with open black circles, intact complement with closed black circles, and C1q-depleted complement with open grey circles. Data are presented as fold increase in the mean number of adherent monocytes for each stimulation, normalized to untreated under each respective complement condition. Paired measurements within the same experiment are indicated with lines connecting the scatter dots. Groups were compared by 2-way ANOVA (repeated measures) followed by Bonferroni multiple comparisons test. ns P > 0.05, *P < 0.05 compared to intact complement. ANOVA., analysis of variance.

FIGURE 2

Monocytic cell adhesion to monoclonal HLA I antibody- and human HLA alloserum-stimulated endothelial cells is significantly enhanced in the presence of activated human complement. HAEC were left untreated, treated with negative control human IgG1 (control IgG) which does not bind to HLA or endothelial cells, monoclonal HLA class I antibodies at 100 ng/mL, negative human serum (NS) or HLA allosera for 20 minutes. The human monocytic cell line MM6 was fluorescently labeled with CFSE, then added to endothelial monolayers in the presence of 25% heat-inactivated human complement (HI C’, black border) or intact complement (25% C’, dark grey fill) and allowed to adhere for 45 minutes. Nonadherent MM6 were removed by washing and monolayers were fixed. Adherent MM6 were visualized by fluorescence microscopy in the FITC channel, and counted in 5 to 8 fields per condition. A, Box and whiskers plot shows 1 representative experiment of MM6 binding to HAEC from donor 3F1153 treated with chimeric pan HLA class I hIgG1 (HLA I IgG1) or human allele-specific HLA-A2/A28 IgG1, which recognizes HLA-A2 carried by these endothelial cells. The number of adherent MM6 per field is given with error bars demonstrating the range from minimum to maximum, and central line at the median. Dashed line represents the baseline MM6 adherence in the untreated condition with HI C'. B, Scatter dots summarize results from multiple independent experiments (control IgG, n = 4; HLA IgG, n = 9 combinations of HAEC from 3 different donors stimulated with monoclonal HLA I IgG1 and relevant HLA-A IgG1). Each data point represents a unique HAEC-antibody combination, that is, H126 stimulated with HLA-A3/A11 IgG1 or 3F1153 treated with HLA-A2/A28 IgG1. Groups were compared by 2-way ANOVA followed by Bonferroni multiple comparisons test. ns, P > 0.05; ***P ≤ 0.001 comparing HI C’ to 25% C’ for each stimulation. C, Box and whiskers plot shows 1 representative experiment of MM6 binding to HAEC from donor X127 stimulated with negative serum (NS) or human alloserum J1179, which recognizes HLA-A24 carried by these cells. The number of adherent MM6 per field is given with error bars demonstrating the range from minimum to maximum, and central line at the median. Dashed line represents the baseline MM6 adherence in the untreated condition with HI C’. D, Scatter dots summarize results from multiple independent experiments (NS, n = 4; alloserum, n = 6 different allosera against relevant HAEC). Groups were compared by 2-way ANOVA followed by Bonferroni multiple comparisons test. ns, P > 0.05; ***P ≤ 0.001 comparing HI C' to 25% C' for each stimulation.

FIGURE 3

A specific inhibitor of the classical complement pathway, TNT003, significantly dampens complement-mediated adhesion of monocytic cells to HLA antibody-activated endothelial cells. HAEC were left untreated, treated with negative control human IgG1 (control IgG) which does not bind to HLA or endothelial cells, monoclonal HLA class I antibodies at 100 ng/mL, negative human serum (NS) or HLA allosera for 20 minutes. The human monocytic cell line MM6 was fluorescently labeled with CFSE, then added to endothelial monolayers in the presence of intact human complement preincubated with 25 μg/mL of isotype control F(ab′)₂ fragment of mIgG1 (CTR), or with TNT003 F(ab′)₂ fragment at increasing concentrations. As above, adherent MM6 were visualized by fluorescence microscopy in the FITC channel, and counted in 5 to 8 fields per condition. A, Box and whiskers plot illustrates results from 1 representative experiment of HAEC H126 activated with anti-HLA-A3/A11 IgG1, which recognizes HLA-A3 expressed by these cells. The number of adherent MM6 per field is given with error bars demonstrating the range from minimum to maximum, and central line at the median. Dashed line represents the baseline MM6 adherence in the untreated condition (no HLA Ab). B, Box and whiskers plot illustrates results from 1 representative experiment of HAEC H126 activated with pooled positive serum (PS), which recognizes HLA-A3 and A29 expressed by these cells. The number of adherent MM6 per field is given with error bars demonstrating the range from minimum to maximum, and central line at the median. Dashed line represents the baseline MM6 adherence in the untreated condition (no HLA Ab). C, Scatter dots illustrate the percent inhibition of MM6 adherence by TNT003 at different concentrations (1-25 μg/mL) in 4 independent experiments (alloserum, n = 2; monoclonal HLA antibody, n = 2). Line indicates the median. D, Scatter dots summarize results from multiple independent experiments measuring MM6 adhesion to HAEC stimulated with monoclonal HLA class I antibodies in the presence of complement with a negative control antibody or with TNT003 at 10 μg/mL (control IgG, n = 2; HLA IgG1, n = 7). Groups were compared by 2-way ANOVA followed by Bonferroni multiple comparisons test. ns, P > 0.05, **P ≤ 0.01 comparing control-treated 25% C' to TNT003 for each stimulation. E, Scatter dots summarize results from multiple independent experiments measuring MM6 adhesion to HAEC stimulated with polyclonal allosera in the presence of complement with a negative control antibody or with TNT003 at 10 μg/mL (NS, n = 2; alloserum, n = 6). Groups were compared by 2-way ANOVA followed by Bonferroni multiple comparisons test. ns, P > 0.05, ***P ≤ 0.001 comparing control-treated 25% C’ to TNT003 for each stimulation. F, Scatter dots illustrate the percent inhibition of MM6 adherence by TNT003 at 10 μg/mL over multiple independent experiments (HLA monoclonal antibody mAb, n = 8; alloserum, n = 8). Line indicates the median percent inhibition.

FIGURE 4

A specific inhibitor of the classical complement pathway, TNT003, and antibody to C5 significantly dampen complement-mediated adhesion of primary monocytes to HLA alloserum-activated endothelial cells. HAEC were left untreated, or stimulated with negative serum (NS) or relevant HLA allosera for 20 minutes. Peripheral blood monocytes (PBMC-mo) were added in the presence of intact human complement with control F(ab′)₂ (dark grey fill), anti-C5 mIgG1 (grey border), or TNT003 F(ab′)₂ (light grey fill) at 10 μg/mL and allowed to adhere for 45 minutes. Adherent monocytes were counted in 5 to 8 fluorescent fields per condition. A, Box and whiskers plot from 1 representative experiment of HAEC 3 F1153 stimulated with allosera M8177 and M4277, both recognizing HLA-A2 expressed by these endothelial cells. The number of adherent monocytes per field is given with error bars demonstrating the range from minimum to maximum, and central line at the median. Dashed line represents the baseline monocyte adherence in the untreated condition with control F(ab′)₂-treated complement. B, Scatter dots summarize results from multiple independent measurements of peripheral blood monocyte adhesion to HAEC from different donors (NS, n = 7; alloserum, n = 14). Fold increase was calculated by dividing the mean number of adherent monocytes in NS and allosera stimulation by untreated under each respective complement condition. Line indicates mean fold increase in each group. Groups were compared using 2-way ANOVA followed by Tukey multiple comparisons test. ns, P > 0.1; ^#P < 0.1; ***P < 0.001 compared with Control F(ab’)₂. C, Scatter dots illustrate the percent inhibition of monocyte adherence by anti-C5 and TNT003 at 10 μg/mL in 4 independent experiments (n = 14 for both, 3 different monocyte donors against 2 different HAEC with multiple allosera). Each data point represents a unique endothelium-serum-monocyte combination. Line indicates the median. D, Line graph shows the same data as in (B) with mean fold increase of monocyte adherence ±SEM (NS, n = 7; n = 14 EC-alloserum-monocyte combinations). Groups were compared using 2-way ANOVA followed by Tukey multiple comparisons test. *P < 0.01 comparing allosera and anti-C5 versus TNT003.

FIGURE 5

Endothelial induction of P-selectin by HLA I antibodies is enhanced in the presence of complement. HAEC were left untreated or stimulated for 20 minutes with negative control IgG1 (control IgG), monoclonal HLA antibodies, allosera or PMA as a positive control. Heat-inactivated complement (open circles) or intact complement (closed circles) was added at the beginning of stimulation to a final concentration of 25%. Monolayers were fixed and cell surface P-selectin was probed on unpermeabilized cells by cell-based ELISA. A, Representative experiment showing induction of cell surface P-selectin on HAEC from donor Y126, stimulated with anti-HLA-A3/A11 IgG1 at indicated concentrations or relevant allosera J1179 and M0546. Open circles represent heat-inactivated complement (HI C'); closed black circles represent intact complement. PMA was used as a biological positive control for WPB exocytosis. Cells permeabilized with Triton X 100 were used as a technical positive control for detection of P-selectin (both cell surface and intracellular). Results are presented as the mean OD of P-selectin staining from technical triplicate well measurements with error bars indicating the range. Dashed line indicates basal P-selectin detected on the surface of untreated endothelial cells. B, Scatter dots illustrate the results from multiple independent measurements of cell surface P-selectin by HLA antibodies, in the presence of heat-inactivated (open circles) or intact complement (closed black circles). n = 9 for control IgG 100 ng/mL and PMA 200 nM, n = 11 for anti-HLA IgG1 at 100 ng/mL. HI C' was compared to intact C’ for each stimulation using 2-way ANOVA followed by Bonferroni multiple comparisons test. *P < 0.05, ****P < 0.0001 comparing HI C’ to intact C'. C, Scatter dots illustrate the results from multiple independent measurements of cell surface P-selectin by HLA allosera, in the presence of heat-inactivated (white dots) or intact complement (black dots). n = 5 for negative serum NS, n = 17 alloserum-EC combinations. Statistical differences between groups were compared using 2-way ANOVA followed by Bonferroni multiple comparisons test. ns P > 0.1; ****P < 0.0001 comparing HI C' to intact C'. D, Scatter dots represent the results from multiple independent measurements of cell surface P-selectin induced by monoclonal HLA IgG1 or HLA allosera in the presence of intact complement (black circles) or C1q-depleted complement (grey squares) (control IgG and PMA, n = 3; HLA IgG1, n = 6; allosera, n = 6). Groups were compared using 2-way ANOVA followed by uncorrected Fisher least significant difference (LSD) test. ^#P < 0.1; ***P < 0.001 comparing intact complement to C1q-depleted. OD, optical density.

FIGURE 6

Stimulation of endothelial cells with complement anaphylatoxins triggers rapid P-selectin expression and increased adhesion of monocytic cells. A, HAEC were left untreated or stimulated with purified C3a (grey circles) or C5a (black circles) at 1 to 100 nM for 5 minutes, or PMA at 200 nM for 20 minutes. Endothelial monolayers were fixed and cell surface P-selectin was measured by cell-based ELISA. A representative experiment shows HAEC from donor X127, with mean OD ±range in technical duplicate wells. B, HAEC were left untreated or stimulated with purified C3a (grey circles) or C5a (black circles) at 1 to 100 nM for 5 minutes, or PMA at 200 nM for 20 minutes. Scatter dots represent fold increase in cell surface P-selectin normalized to untreated conditions over multiple independent experiments (n = 6). Line indicates the median fold increase. Statistical differences between groups were compared using 1-way ANOVA followed by Bonferroni multiple comparisons test. ns P > 0.05; **P < 0.01; ***P < 0.001, ****P < 0.0001 compared with untreated. C, HAEC from donor H126 were left untreated or stimulated with purified C3a (grey bars) or C5a (black bars) at 1 to 100 nM for 5 minutes. MM6 were added to endothelial monolayers and allowed to adhere for 20 minutes. Box and whiskers plot of 1 representative experiment shows the number of adherent MM6 per field, with error bars from minimum to maximum, and line at the median. Dashed line indicates basal MM6 adhesion to untreated endothelial cells. D, HAEC were left untreated or stimulated with purified C3a (grey circles) or C5a (black circles) at 1 to 100 nM for 5 minutes. MM6 were added to endothelial monolayers and allowed to adhere for 20 minutes. Scatter dots represent results from multiple independent experiments (n = 3). Statistical differences between groups were compared using 2 way ANOVA followed by Bonferroni multiple comparisons test. **P < 0.01; ***P < 0.001; ****P < 0.0001 compared to untreated (0).

FIGURE 7

HLA antibodies and complement anaphylatoxins additively increase endothelial P-selectin and adhesion of monocytic cells. HAEC were left untreated or stimulated with chimeric HLA I IgG1 at 100 ng/mL for 20 minutes, C3a 10 nM or C5a 10 nM alone for 5 minutes, or HLA IgG1 for 20 minutes in combination with C3a or C5a for the final 5 minutes of stimulation. A, Representative experiment showing induction of cell surface P-selectin on HAEC from donor 3F1153, stimulated with C3a or C5a, with (black circles) or without (open circles) anti-HLA I IgG1 at 100 ng/mL. The mean optical density (OD) of cell surface P-selectin +/− range is given for duplicate wells. Dashed line indicates basal P-selectin cell surface expression on untreated endothelial cells. B, Scatter dots summarize multiple experiments (n = 4) comparing P-selectin induction on HAEC with C3a alone, C5a alone (open circles), HLA IgG1 alone, or C3a and C5a in combination with HLA antibody at 100 ng/mL (black circles). Statistical differences between groups were compared using 2 way ANOVA followed by Bonferroni multiple comparisons test. **P < 0.01; ****P < 0.0001 comparing no HLA antibody to with HLA antibody. † P < 0.01 comparing C3a or C5a with HLA IgG1 to HLA IgG1 with no complement added. C, Box and whiskers plot illustrates results from a representative experiment of MM6 adhesion to HAEC from donor 3F1153 stimulated with C3a or C5a alone (grey bars) or in combination with HLA I IgG1 (black bars). Dashed line indicates background MM6 adhesion to untreated endothelial cells. Solid lines indicate paired measurements within the same experiment. D, Scatter dots summarize multiple experiments (n = 3) comparing MM6 adhesion to endothelial cells stimulated with C3a or C5a alone (open circles) or in combination with HLA IgG1 at 100 ng/mL (black circles). Dashed line indicates background MM6 adhesion to untreated endothelial cells. Solid lines indicate paired measurements within the same experiment.

FIGURE 8

Stimulation of monocytes with complement anaphylatoxins causes increased adhesion to endothelial cells. MM6 were left untreated or stimulated directly with C3a (grey) or C5a (black) at different concentrations (1-100 nM). Then, MM6 were added to untreated endothelial cell monolayers and allowed to bind for 30 minutes. Adherent MM6 were counted by fluorescence microscopy. A, Box and whiskers plot shows 1 representative experiment of C3a or C5a-stimulated MM6 binding to HAEC from donor H126. Results are presented as number of adherent monocytes per field, with error bars from minimum to maximum and line at the median. Dashed line indicates background MM6 adhesion to untreated endothelial cells. B, Scatter dots illustrate results from multiple experiments (n = 5) showing MM6 adhesion to HAEC after stimulation with C3a or C5a at 1 to 100 nM. Groups were compared by 2 way ANOVA followed by Bonferroni multiple comparisons test. ^#P < 0.1; **P < 0.01; ***P < 0.001 compared to untreated (0). C, MM6 were pretreated with neutralizing antibodies to CD11a (LFA-1), CD11b (Mac-1), CD29 (VLA-4), or an isotype negative control at 10 μg/mL and stimulated directly with C3a or C5a at 1nM. Then, MM6 were added to unstimulated HAEC and allowed to adhere for 20 minutes. Box and whiskers plot shows the number of adherent MM6 per field in 1 representative experiment with HAEC donor 3F1153, with error bars from minimum to maximum and line at the median. Dashed line indicates basal adhesion of untreated MM6 to untreated HAEC. D, Summary of 3 independent experiments measuring inhibition of C3a (grey circles) and C5a (black circles)-induced MM6 adhesion to HAEC by blocking CD11a or CD11b. Groups were compared by 2 way ANOVA followed by Bonferroni multiple comparisons test. *P < 0.05 compared with no inhibitor.